You are on page 1of 300

Information

System

GPRS/EGPRS Global Description

A30808-X3247-L24-5-7618
GPRS/EGPRS Global Description Information
System

f Important Notice on Product Safety


DANGER - RISK OF ELECTRICAL SHOCK OR DEATH – FOLLOW ALL INSTALLATION INSTRUCTIONS.
The system complies with the standard EN 60950 / IEC 60950. All equipment connected to the system must
comply with the applicable safety standards.
Hazardous voltages are present at the AC power supply lines in this electrical equipment. Some components may
also have high operating temperatures.
Failure to observe and follow all installation and safety instructions can result in serious personal injury
or property damage.
Therefore, only trained and qualified personnel may install and maintain the system.

The same text in German:


Wichtiger Hinweis zur Produktsicherheit
LEBENSGEFAHR - BEACHTEN SIE ALLE INSTALLATIONSHINWEISE.
Das System entspricht den Anforderungen der EN 60950 / IEC 60950. Alle an das System angeschlossenen
Geräte müssen die zutreffenden Sicherheitsbestimmungen erfüllen.
In diesen Anlagen stehen die Netzversorgungsleitungen unter gefährlicher Spannung. Einige Komponenten
können auch eine hohe Betriebstemperatur aufweisen.
Nichtbeachtung der Installations- und Sicherheitshinweise kann zu schweren Körperverletzungen oder
Sachschäden führen.
Deshalb darf nur geschultes und qualifiziertes Personal das System installieren und warten.

Caution:
This equipment has been tested and found to comply with EN 301489. Its class of conformity is defined in table
A30808-X3247-X910-*-7618, which is shipped with each product. This class also corresponds to the limits for a
Class A digital device, pursuant to part 15 of the FCC Rules.
These limits are designed to provide reasonable protection against harmful interference when the equipment is
operated in a commercial environment.
This equipment generates, uses and can radiate radio frequency energy and, if not installed and used in accor-
dance with the relevant standards referenced in the manual “Guide to Documentation”, may cause harmful inter-
ference to radio communications.
For system installations it is strictly required to choose all installation sites according to national and local require-
ments concerning construction rules and static load capacities of buildings and roofs.
For all sites, in particular in residential areas it is mandatory to observe all respectively applicable electromagnetic
field / force (EMF) limits. Otherwise harmful personal interference is possible.

Trademarks:
All designations used in this document can be trademarks, the use of which by third parties for their own purposes
could violate the rights of their owners.

Copyright (C) Siemens AG 2004.

Issued by the Information and Communication Mobile Group


Hofmannstraße 51
D-81359 München

Technical modifications possible.


Technical specifications and features are binding only insofar as
they are specifically and expressly agreed upon in a written contract.

2 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

This document consists of a total of 300 Pages. All pages are issue 5.

Reason for Update


Chapter/Section Reason for Update

All New Release BR7.0.

Issue History
Issue Date of Reason for Update
Number Issue

1 07/2003 First issue for the New Release BR7.0.


2 09/2003 Second issue for BR7.0
3 12/2003 Third issue for BR7.0
4 03/2004 Fourth issue for BR7.0
5 08/2004 Fifth issue for BR7.0

A30808-X3247-L24-5-7618 3
GPRS/EGPRS Global Description Information
System

Contents
1 Introductions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.1 Generality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2 Structure of the Manual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2 Siemens Features Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15


2.1 BR5.5 Feature Description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 BR6.0 Feature Description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3 BR7.0 Feature Description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3 GPRS/EGPRS Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.1 GPRS and EGPRS Modulation Principles . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.2 Network Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.3 GPRS/EGPRS Protocol Stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.4 Data Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.5 RLC/MAC Block and Radio Block Structures. . . . . . . . . . . . . . . . . . . . . . . . 40
3.5.1 RLC/MAC and Radio Block Structures: Data Transfer . . . . . . . . . . . . . . . . 40
3.5.1.1 RLC/MAC Block and Radio Block Structures for GPRS Data Transfer . . . . 40
3.5.1.2 RLC/MAC Block and Radio Block Structure for EGPRS Data Transfer. . . . 41
3.5.2 RLC/MAC Block Structure: Control Signalling . . . . . . . . . . . . . . . . . . . . . . . 42

4 Radio Interface Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44


4.1 GPRS/EGPRS Physical Channels. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.2 Channel Coding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.2.1 GPRS Channel Coding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.2.2 EGPRS Channel Coding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.3 Temporary Block Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.3.1 Multiplexing MSs on the same PDCH: Downlink Direction . . . . . . . . . . . . . 56
4.3.2 Multiplexing MSs on the same PDCH: Uplink Direction. . . . . . . . . . . . . . . . 57
4.3.3 Multiplexing MSs on the same PDCH: Configuration. . . . . . . . . . . . . . . . . . 59
4.4 GPRS/EGPRS Logical Channels. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.4.1 Packet Broadcast Control Channel (PBCCH) . . . . . . . . . . . . . . . . . . . . . . . 60
4.4.2 Packet Common Control Channel (PCCCH) . . . . . . . . . . . . . . . . . . . . . . . . 61
4.4.3 Packet Data Traffic Channel (PDTCH) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.4.4 Packet Dedicated Control Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.4.5 Coding of GPRS/EGPRS Logical Channels . . . . . . . . . . . . . . . . . . . . . . . . 64
4.5 Mapping of Logical Channels onto Physical Channels . . . . . . . . . . . . . . . . 64
4.5.1 PDCH without the Specific GPRS/EGPRS Signalling . . . . . . . . . . . . . . . . . 65
4.5.2 PDCH Carrying both PBCCH and PCCCH . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.5.3 PDCH Carrying PCCCH. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.6 Packet Timing Advance Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.6.1 Initial Timing Advance Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.6.2 Continuous Timing Advance Update . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.7 Multislot Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.7.1 Mobile Station Classes for Multislot Capabilities . . . . . . . . . . . . . . . . . . . . . 71
4.7.2 Mapping of Uplink Packet Traffic Logical Channels. . . . . . . . . . . . . . . . . . . 72
4.7.3 Mapping of Downlink Packet Traffic Logical Channels . . . . . . . . . . . . . . . . 73

4 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

5 Radio Resources Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74


5.1 Enabling Packet Switched Services in a Cell . . . . . . . . . . . . . . . . . . . . . . . 75
5.1.1 Enabling GPRS Service in the Cell. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.1.2 Enabling EGPRS Service in the Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
5.1.3 Aspects Related to Carrier Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.2 Configuration of GPRS Channels in a Cell . . . . . . . . . . . . . . . . . . . . . . . . . 82
5.3 Management of Packet Data Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5.3.1 Generalities about Resource Assignments. . . . . . . . . . . . . . . . . . . . . . . . . 85
5.3.2 Horizontal/Vertical Allocation Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
5.3.2.1 Vertical Allocation Strategy (VA). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
5.3.2.2 Horizontal Allocation Strategy (HA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.3.2.3 Switching between VA and HA According to Radio Conditions . . . . . . . . . 88
5.3.2.4 Switching between VA and HA according to Abis Interface Conditions . . . 91
5.3.2.5 Allocation of Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
5.3.3 Management of Incoming GPRS/EGPRS Requests . . . . . . . . . . . . . . . . . 92
5.3.3.1 PCU Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.3.3.2 TDPC Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
5.3.4 Upgrading Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.3.4.1 Upgrade of Radio Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.3.4.2 Upgrade of Abis Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.3.5 Incoming CS Calls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.3.6 Waiting Queue Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.3.6.1 Pre-emption of PDCH Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.3.6.2 Pre-emption of PDT Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.3.6.3 Forced Intracell Handovers of Already Established CS Calls . . . . . . . . . . 112

6 Hardware and Software Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113


6.1 Supported BSC Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
6.1.1 “Standard” BSC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
6.1.2 High Capacity BSC with the Old Rack . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
6.1.3 High Capacity BSC with the New Rack . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.1.4 PPCU and PPXU Redundancy and Configuration Rules . . . . . . . . . . . . . 122
6.2 BTS Equipment Supporting GPRS and EGPRS. . . . . . . . . . . . . . . . . . . . 123
6.3 PCU Frames and Dynamic Allocation on the Abis Interface. . . . . . . . . . . 124
6.3.1 Concatenated PCU Frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
6.3.2 Hardware supporting Flexible Abis Allocation and Concatenated PCU Frames
129
6.3.3 Configuration of the Abis Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
6.3.4 Algorithms Regarding Flexible Abis Allocation . . . . . . . . . . . . . . . . . . . . . 133
6.3.5 Abis over satellite links . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
6.4 Packet Switched Services Supported on CCCH/PCCCH . . . . . . . . . . . . . 135

7 Gb Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
7.1 Physical Layer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
7.2 Network Service Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
7.2.1 Sub-Network Service: Frame Relay on Gb Interface . . . . . . . . . . . . . . . . 144
7.2.1.1 Examples of Addressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
7.2.1.2 Frame Relay Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

A30808-X3247-L24-5-7618 5
GPRS/EGPRS Global Description Information
System

7.2.1.3 Procedures for PVCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154


7.2.2 Network Service Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
7.2.2.1 Load Sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
7.2.2.2 Control Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
7.3 BSSGP Protocol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
7.3.1 BSSGP Addressing: BSSGP Virtual Connections (BVCs). . . . . . . . . . . . . 158
7.3.1.1 BVC Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
7.3.2 Quality of Service (QoS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
7.3.3 SGSN-BSS Flow Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
7.3.3.1 MS Flow Control Message . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
7.3.3.2 BVC Flow Control Message . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
7.3.3.3 Flow Control sending criteria (for both BVC and MS) . . . . . . . . . . . . . . . . 172

8 Load Control for Packet Switched Services . . . . . . . . . . . . . . . . . . . . . . . . 175


8.1 Dynamic PTPPKF Reconfiguration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
8.1.1 System Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
8.1.2 Creation of a PCU Object and Enabling a NSVC for It . . . . . . . . . . . . . . . 178
8.1.3 PCU Crash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
8.1.4 PCU Comes Back in Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
8.1.5 Time Needed to Execute PTPPKF Reconfiguration . . . . . . . . . . . . . . . . . 184
8.2 PCU Overload Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

9 GPRS/EGPRS Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186


9.1 Mobile Stations for Packet Switched Services . . . . . . . . . . . . . . . . . . . . . . 186
9.2 Network Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
9.3 Mobility Management Functionalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
9.3.1 Mobility Management States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
9.3.1.1 IDLE State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
9.3.1.2 STAND-BY State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
9.3.1.3 READY State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
9.3.2 Mobility Management Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
9.3.2.1 Attach Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
9.3.2.2 Detach Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
9.4 Radio Resource Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
9.4.1 Packet Idle State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
9.4.2 Packet Transfer State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
9.5 Correspondence between RR States and MM States . . . . . . . . . . . . . . . . 193
9.6 Packet Data Protocol Functionalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
9.6.1 INACTIVE State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
9.6.2 ACTIVE State. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
9.7 Activation and Deactivation of a PDP Context . . . . . . . . . . . . . . . . . . . . . . 195
9.7.1 PDP Context Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
9.7.2 PDP Context Deactivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
9.8 Access to the Network (Establishment of a TBF). . . . . . . . . . . . . . . . . . . . 196
9.8.1 Medium Access Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
9.8.2 TBF Establishment Initiated by the MS on CCCH/PCCCH . . . . . . . . . . . . 197
9.8.2.1 8 Bit or 11 Bit Uplink Access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
9.8.2.2 Establishment using a One Phase Access . . . . . . . . . . . . . . . . . . . . . . . . 199

6 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

9.8.2.3 TBF Establishment using a Two Phases Access . . . . . . . . . . . . . . . . . . . 200


9.8.2.4 TBF Establishment for EDGE Mobile Stations . . . . . . . . . . . . . . . . . . . . . 202
9.8.2.5 Contention Resolution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
9.8.2.6 Uplink Access on PRACH (Access Persistence Control) . . . . . . . . . . . . . 204
9.8.3 TBF Establishment Initiated by the Network on CCCH/PCCCH . . . . . . . . 205
9.8.3.1 Network Operation Modes for Paging. . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
9.8.3.2 Discontinuous Reception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
9.8.4 Relative Reserved Block Period Field (RRBP) . . . . . . . . . . . . . . . . . . . . . 211
9.8.5 Polling Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
9.9 RLC Data Block Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
9.9.1 Acknowledged Mode for RLC/MAC Operation . . . . . . . . . . . . . . . . . . . . . 213
9.9.1.1 GPRS Acknowledged Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
9.9.1.2 EGPRS Acknowledged Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
9.9.2 Unacknowledged Mode for RLC/MAC Operation . . . . . . . . . . . . . . . . . . . 216
9.9.3 Operations on Uplink TBF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
9.9.3.1 Uplink TBF Using the Acknowledged Mode . . . . . . . . . . . . . . . . . . . . . . . 216
9.9.3.2 Uplink TBF Using the Unacknowledged Mode . . . . . . . . . . . . . . . . . . . . . 218
9.9.3.3 Anomalies During an Uplink TBF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
9.9.3.4 Release of an Uplink TBF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
9.9.4 Operations on Downlink TBF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
9.9.4.1 Acknowledged and Unacknowledged Modes on Downlink TBFs . . . . . . . 222
9.9.4.2 Release of a Downlink TBF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
9.9.5 Notes About Concurrent TBFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
9.9.6 Suspend/Resume Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
9.9.7 Notes About GPRS/EGPRS TBF Scheduling. . . . . . . . . . . . . . . . . . . . . . 230
9.9.7.1 Supported QoS Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
9.9.7.2 Scheduling Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

10 GPRS/EGPRS Functionalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234


10.1 Cell Selection and Re-selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
10.1.1 Measurements for Cell Selection and Re-selection . . . . . . . . . . . . . . . . . 234
10.1.2 Cell selection and Re-selection Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . 236
10.1.2.1 GPRS/EGPRS Path Loss Criterion (C1 Criterion) . . . . . . . . . . . . . . . . . . 236
10.1.2.2 C31 Criterion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
10.1.2.3 C32 Criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
10.1.3 Cell Re-selection Algorithm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
10.1.4 Management of GPRS/EGPRS Neighboring Cells. . . . . . . . . . . . . . . . . . 242
10.1.4.1 Handling of Neighboring Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
10.1.4.2 GPRS/EGPRS Neighboring Cells and Involved Parameters . . . . . . . . . . 244
10.1.4.3 Configuration of an Adjacent Cell with GSUP= TRUE . . . . . . . . . . . . . . . 245
10.1.4.4 Configuration of an Adjacent Cell with GSUP= FALSE . . . . . . . . . . . . . . 246
10.1.5 Abnormal Cell Re-selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
10.2 Cell Re-selection from GSM/GPRS/EGPRS Network to UMTS Network . 248
10.2.1 GSM-UMTS Re-selection Algorithm: Circuit Switched Case . . . . . . . . . . 248
10.2.2 GSM-UMTS Re-selection Algorithm: Packet Switched Case . . . . . . . . . . 249
10.2.3 Handling of UMTS Neighboring Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
10.3 Network Controlled Cell Reselection and Traffic Control Management . . 252
10.3.1 Network Controlled Cell Reselection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

A30808-X3247-L24-5-7618 7
GPRS/EGPRS Global Description Information
System

10.3.1.1 Measurement Reporting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255


10.3.1.2 Radio Link Network Controlled Cell Reselection Algorithm . . . . . . . . . . . . 256
10.3.2 GPRS/EGPRS Traffic Control Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . 258
10.3.2.1 Network Controlled Cell Reselection Algorithm for Traffic Control Strategy . .
259
10.4 Power Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
10.4.1 Power Control Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
10.4.2 Measurement at the MS Side. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
10.4.2.1 Packet Idle Mode: Measurements for Power Control. . . . . . . . . . . . . . . . . 263
10.4.2.2 Packet Transfer Mode: Measurements for Power Control . . . . . . . . . . . . . 264
10.4.2.3 Derivation of Channel Quality Reports. . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
10.4.3 BTS Output Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
10.5 Link Adaptation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
10.5.1 Link Adaptation for GPRS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
10.5.1.1 GPRS: Switching Points. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
10.5.1.2 “Quality Traps” Disadvantage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
10.5.1.3 GPRS: Link Adaptation Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
10.5.2 Link Adaptation for EGPRS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
10.5.2.1 EGPRS: Switching Points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
10.5.2.2 EGPRS: Link Adaptation Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
10.5.3 Selection of the Candidate Initial Coding Scheme . . . . . . . . . . . . . . . . . . . 280

11 Database Parameters and Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283

12 Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298

8 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Illustrations
Fig. 3.1 Basic GMSK Constellation of Signal Vectors. . . . . . . . . . . . . . . . . . . . . 30
Fig. 3.2 Basic 8 PSK Constellation of Signal Vectors . . . . . . . . . . . . . . . . . . . . . 31
Fig. 3.3 GPRS/EGPRS Network Architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Fig. 3.4 Protocol Stack for Data Transmission in GPRS/EGPRS Network. . . . . 35
Fig. 3.5 Data Flow across Protocol Layers in case of GPRS/
EGPRS(MSC1...MSC6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Fig. 3.6 Data Flow across Protocol Layers in case of EGPRS(MSC7...MSC9) . 37
Fig. 3.7 Data Flow from the SGSN to the MS. . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Fig. 3.8 RLC/MAC block’s structure for Data Transfer . . . . . . . . . . . . . . . . . . . . 40
Fig. 3.9 Radio Block structure for Data Transfer on the “Um” Interface . . . . . . . 40
Fig. 3.10 RLC/MAC Block structure for Data Transfer with one RLC Data Block field
41
Fig. 3.11 RLC/MAC Block structure for Data Transfer with two RLC Data block fields
41
Fig. 3.12 Radio Block for Data Transfer with one RLC Data Block field . . . . . . . . 42
Fig. 3.13 Radio Block for Data Transfer with two RLC Data Block field . . . . . . . . 42
Fig. 3.14 RLC/MAC Block Structure for Control Messages . . . . . . . . . . . . . . . . . 42
Fig. 3.15 Radio Block for Control Messages (Signalling).. . . . . . . . . . . . . . . . . . . 42
Fig. 4.1 Packet Data Channel (PDCH) within a TDMA frame. . . . . . . . . . . . . . . 45
Fig. 4.2 Multiframe Structure for a PDCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Fig. 4.3 GPRS Coding Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Fig. 4.4 Coding of the RLC/MAC Block using CS-1 . . . . . . . . . . . . . . . . . . . . . . 50
Fig. 4.5 EGPRS Coding Schemes and Families. . . . . . . . . . . . . . . . . . . . . . . . . 52
Fig. 4.6 Interleaving of MCS9 Coded Data into Two Consecutive Normal Bursts54
Fig. 4.7 Interleaving of MCS6 Coded Data into Four Consecutive Normal Bursts .
55
Fig. 4.8 Multiplexing Mobile Station on the same PDCH (Downlink) . . . . . . . . . 57
Fig. 4.9 Multiplexing Mobile Station on the same PDCH (Uplink). . . . . . . . . . . . 58
Fig. 4.10 Example of Mapping of the PBCCH Channel. . . . . . . . . . . . . . . . . . . . . 61
Fig. 4.11 Packet Common Control Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Fig. 4.12 Example of Mapping of the PCCCH Channel. . . . . . . . . . . . . . . . . . . . . 62
Fig. 4.13 Example of Mapping of two PCCCH Channels.. . . . . . . . . . . . . . . . . . . 63
Fig. 4.14 Example of Mapping of Logical Channels in the Physical Channel (Down-
link Direction). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Fig. 4.15 Example of Mapping of Logical Channels in the Physical Channel (Uplink
Direction).. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Fig. 4.16 Example of Downlink Configuration with PBCCH and PCCCH Channels .
67
Fig. 4.17 Example of Uplink Configuration with PRACH Channel. . . . . . . . . . . . . 68
Fig. 4.18 Continuous Timing Advance Update Feature . . . . . . . . . . . . . . . . . . . . 70
Fig. 4.19 Example of Multislot Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Fig. 5.1 Example of TRXs enabled to support Packet Switched Services.. . . . . 77
Fig. 5.2 Example of TRXs enabled to support GPRS and EGPRS. . . . . . . . . . . 80
Fig. 5.3 Example of GPRS/EGPRS configuration. . . . . . . . . . . . . . . . . . . . . . . . 84

A30808-X3247-L24-5-7618 9
GPRS/EGPRS Global Description Information
System

Fig. 5.4 Example of Vertical Allocation Algorithm . . . . . . . . . . . . . . . . . . . . . . . . 87


Fig. 5.5 Example of Horizontal Allocation Algorithm . . . . . . . . . . . . . . . . . . . . . . 88
Fig. 5.6 Example of a Cell Configured with Five TRXs. . . . . . . . . . . . . . . . . . . . . 89
Fig. 5.7 Allocation Algorithm followed by the PCU. . . . . . . . . . . . . . . . . . . . . . . . 99
Fig. 5.8 Allocation Algorithm followed by the TDPC . . . . . . . . . . . . . . . . . . . . . 104
Fig. 6.1 Hardware and Software Entities supporting the GPRS/EGPRS technology
113
Fig. 6.2 View of the BSC Rack with and without PPCU Boards. . . . . . . . . . . . . 116
Fig. 6.3 View of the “High Capacity” BSC with the Traditional Rack.. . . . . . . . . 118
Fig. 6.4 High Capacity BSC with the New Rack . . . . . . . . . . . . . . . . . . . . . . . . 122
Fig. 6.5 Fundamental Principle of Concatenated PCU Frames . . . . . . . . . . . . . 126
Fig. 6.6 Abis Mapping for a downlink MCS9 radio block requiring 5 Abis subslots .
128
Fig. 6.7 High Capacity BSC: Relationship between PCU Frames and Abis Alloca-
tion according to the BTSE Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Fig. 6.8 Standard BSC: Relationship between PCU Frames and Abis Allocation
according to the BTSE Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Fig. 6.9 BSC handling of BTS Equipment with Software Releases not supporting
the Abis Dynamic Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Fig. 6.10 Mapping of CCCH/PCCCH Channels on the Abis Interface. . . . . . . . . 136
Fig. 6.11 CCCH/PCCCH Message Handling. . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Fig. 7.1 Gb Interface: Protocol Stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Fig. 7.2 Different Connection Types between the BSC and the SGSN. . . . . . . 140
Fig. 7.3 Example of Frame Relay Links . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Fig. 7.4 Example of Frame Relay Link (GTS=3). . . . . . . . . . . . . . . . . . . . . . . . . 143
Fig. 7.5 Example of Frame Relay Link (GTS=3&4&5&6). . . . . . . . . . . . . . . . . . 143
Fig. 7.6 Example of Frame Relay Link (GTS=3&4&7&8). . . . . . . . . . . . . . . . . . 143
Fig. 7.7 Network Service Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
Fig. 7.8 Gb Interface with a Frame Relay Network . . . . . . . . . . . . . . . . . . . . . . 145
Fig. 7.9 Creation of a NSVC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Fig. 7.10 BSC Configured with One PCU and Two FR Links (64 kbit/s each). . . 148
Fig. 7.11 BSC Configured with One PCU and Two FR Links (128 kbit/s each one)..
149
Fig. 7.12 BSC Configured with Two PCUs and Two FR Links each one. . . . . . . 150
Fig. 7.13 Frame Relay Network Connecting two DTE Devices . . . . . . . . . . . . . . 152
Fig. 7.14 Frame Relay Frame Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Fig. 7.15 Periodic Polling Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Fig. 7.16 Distribution of Packet Switched Data Traffic among Different Cells . . . 160
Fig. 7.17 Cascaded Flow Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Fig. 7.18 Token Leaky Bucket (in SGSN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
Fig. 7.19 Closed Loop Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Fig. 7.20 Example Cell Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Fig. 7.21 MS-FLOW-CONTROL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Fig. 7.22 SGSN does not answer with MS-FLOW-CONTROL-ACK message . . 173
Fig. 8.1 Example of PTPPKF Distribution During System Initialization . . . . . . . 178

10 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Fig. 8.2 Example of PTPPKF Distribution when a New PCU is Created - Step 1. .
179
Fig. 8.3 Example of PTPPKF Distribution when a New PCU is Created - Step 2. .
180
Fig. 8.4 Example of PTPPKF Distribution in Case of PCU Crash. . . . . . . . . . . 181
Fig. 8.5 Example of PTPPKF Distribution when a PCU Comes Back in Service . .
183
Fig. 9.1 Network Structure: Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
Fig. 9.2 Mobility Management States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
Fig. 9.3 Radio Resource States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Fig. 9.4 Packet Data Protocol States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
Fig. 9.5 Coding of the 11 Bit Access Burst . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
Fig. 9.6 One Phase Access on PCCCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Fig. 9.7 Two Phases Access on CCCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Fig. 9.8 Packet Access Reject Procedure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Fig. 9.9 TBF Establishment Initiated by the Network on PCCCH . . . . . . . . . . . 207
Fig. 9.10 Behavior of T3182 Timer and N3102 Counter . . . . . . . . . . . . . . . . . . . 218
Fig. 9.11 Detection of Anomalies during an Uplink TBF on the Network Side . . 219
Fig. 9.12 Release of an Uplink TBF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Fig. 9.13 Release of Resources on the Network Side during an Uplink TBF (in case
of T3169 timer expiration) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Fig. 9.14 Control Procedure Executed by the Network during a Downlink TBF . 223
Fig. 9.15 Release of a Downlink TBF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
Fig. 9.16 Suspend Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
Fig. 9.17 Resume Procedure (the MS has remained in the same cell - Successful
Resume) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Fig. 9.18 Resume Procedure (The MS has changed the Routing Area) . . . . . . 230
Fig. 10.1 Management of Adjacent Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Fig. 10.2 Network Controlled Cell Reselection Procedure . . . . . . . . . . . . . . . . . 255
Fig. 10.3 CS1 and CS2 Throughput Depending on C/I (dB). . . . . . . . . . . . . . . . 266
Fig. 10.4 Gross Throughput Depending on CS and C/I (dB) . . . . . . . . . . . . . . . 268
Fig. 10.5 BLER as Function of C/I (dB) for all GPRS Coding Schemes . . . . . . . 269
Fig. 10.6 Simulation Results for Family A (+MCS1) without IR . . . . . . . . . . . . . 272

A30808-X3247-L24-5-7618 11
GPRS/EGPRS Global Description Information
System

1 Introductions

1.1 Generality
With the implementation of the second generation of the mobile systems, due to the
digital transmission mode they use, not only pure speech transmission, but also low rate
data transmission and several supplementary services have been provided to the final
users.
Nevertheless, since needs for mobile data transmission are rapidly increasing due to the
current world wide activities based on the exchange of big amount of informations with
the minimum time delay and maximun efficiency the growth in the area of data transmis-
sion is much higher and faster than in the area of speech transmission.
In principle, a higher data transmission rates in the GSM network can be achieved by
the HSCSD feature (High Speed Circuit Switched Data). With HSCSD it is possible to
match the ISDN transmission rate, by combining four timeslots of the TDMA frame.
One disadvantage of the HSCSD feature, however, is the circuit switched data transmis-
sion that it requires and it uses; in fact when circuit switched connections are used the
following limitations arise:
– efficient resource management becomes difficult to reach.
– additional costs arise for the user.
For this reason the HSCSD technology is essentially suited for whose applications
involving high, but constant, transmission rates (e.g., videotelephony).
To further increase data rates, exceeding HSCSD limits, the General Packet Data
Service (GPRS) has been developed.
GPRS is intended to provide the possibility of transmitting large volumes of data in a
very short time; on the other hand it ensures a better management of available
resources, which will:
– increase the number of users;
– reduce the costs arising for individual users (volume-oriented fees).
Using the GPRS technology it is possible to reach a maximum data throughput of about
150-170 kbit/s per each user.
The incoming third generation of mobile networks, however, requires, for its forthcoming
multimedia applications, much more bandwidth, at least 384 kbit/s. The Enhanced
General Packet Data Service (EGPRS) represents the GPRS upgrade and offers the
opportunity to achieve those high data rates by preserving the most important GSM air
interface features (e.g. 200 kHz channeling, TDMA access type, cell planning
processes), by introducing a new modulation scheme (8 PSK instead of GMSK). This
means that EGPRS will rely completely on underlying GSM functionality.
Due to its GSM/GPRS compatibility EGPRS is the optimal packet data feature for estab-
lished GSM operators, it provides a high protection for old investments and requires only
small new investments. Looking at the fact that only a limited number of operators per
country have been assigned UMTS licenses, EGPRS is also a good opportunity for
those operators (so called “UMTS-losers”) to make an evolutionary step to their mobile
networks and provides the opportunity to offer in advance services normally offered by
3rd generation networks.
For this reason it is expected that both UMTS and GPRS/EGPRS networks will coexist
in the near future. UMTS will serve mainly hotspots that require up to 2 Mbit/s data

12 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

services per subscriber and GPRS/EGPRS will be used to cover the rest of the area
offering up to 384 kbit/s data services.

1.2 Structure of the Manual


This manual describes in detail the GPRS/EGPRS technology with a particular attention
to the provided features considering that the satisfaction of almost all customers derives
from an optimized system’s usability and cost saving. Therefore its main purpose is to
allow users in understanding the main characteristics of packet switched (PS) data
services.
Besides the general description of GPRS and EGPRS features the Siemens solution is
detailed; when a subject is shown, the parameters involved in the subject are also
described.
The manual is organized in the following way: when a feature requires the description
of one or more database parameters, each parameter is linked to a specific table of the
BSC:CML manual is executed; that describes:
– the meaning of the parameter;
– the range of the parameter;
– its default value;
– the commands to which the parameter belongs to.
The Chapter: "2 Siemens Features Description"is completely dedicated to identifying all
the Siemens Feature Sheets (or Change Requests) for the GPRS/EGPRS technology.
The following information is shown for each Feature Sheet (or Change Request):
– its number and title;
– a brief description of the Feature Sheet (or Change Request);
– the release of the Feature Sheet (or Change Request).
Finally, in the last chapter of the manual four different tables are inserted:
– in the first table all the parameters, related to the GPRS/EGPRS only, which are
discussed in the manual are listed in the alphabetical order. For each parameter one
or more links to the chapters of the manual where the parameter is described and in
addition also a link to the title of the related Feature Sheets (or Change Requests)
that introduce or describe the parameter in Siemens technology are introduced; in
this way a user, who wants to know the meaning of one parameter, can find in the
manual the location where the parameter is explained and also which are the other
documents that add more information on it;
– in the second table the non specific GPRS/EGPRS parameters which are described
in the manual since they are also related to packet switched services are listed in
alphabetical order. For each parameter, one or more links to the chapters of the
manual where the parameter is described is given. Besides, starting from the param-
eters of the BR5.5 release onwards, a link to the features that describe the param-
eter is assigned.
– in the third table all the database objects which are related to the GPRS/EGPRS
technology only are listed in alphabetical order. For each object the link to the
Feature Sheets (or Change Requests) that introduce or describe the object is given.
– in the fourth table all the non specific GPRS/EGPRS database objects which are
involved in packet switched (PS) services are listed in the alphabetical order.For
each object the link to the chapters of the manual that describe it is given; Besides
starting from the objects of the BR5.5 release onwards, a link to the features that
describe the parameter is also executed.

A30808-X3247-L24-5-7618 13
GPRS/EGPRS Global Description Information
System

The manual is subdivided in several chapters:


• The chapter: "1 Introductions" explains the purpose of the manual and its structure.
A short description is also added to each chapter for introducing the reader to its
content.
• The Chapter:"2 Siemens Features Description" is the collection of all the Siemens
Features related to the GPRS/EGPRS technology; when a parameter is described
in the manual, a link to the feature that is affected is executed;
• The Chapter: "3 GPRS/EGPRS Overview" comprises a general discussion about
the packet switched (PS) services, showing the new architecture, the protocol stack
and the data flow across the several network entities;
• The Chapter:"4 Radio Interface Description" details the GPRS/EGPRS radio inter-
face, that means the configuration of new logical channels, their mapping on phys-
ical channels and the rules that allow:
– sharing the same physical channel among several mobile stations;
– the assignment of more physical channels to the same mobile station;
• The Chapter: "6 Hardware and Software Architecture" describes the hardware and
software modules that are requested for the introduction of packet switched
services; in this chapter also the Packet Control Unit toghether with its main features
are described.
• The Chapter: "5 Radio Resources Management" introduces the not simple concept
of Radio Resource Management and shows how the user can configure the
resources of the cell, to allow him to manage both circuit switched (CS) and packet
switched (PS) services; Toghether with the description are provided some examples
are very important to clarify how the resources can be handled;
• The Chapter: "7 Gb Interface" is dedicated to the Gb interface, i.e., the interface that
connects the BSC to the core network. The frame relay protocol, which character-
izes the Gb interface, is described. In particular, the following topics are detailed:
– the physical layer;
– permanent virtual connections;
– examples of configuration;
– procedures regarding the Gb interface.
• The Chapter: "8 Load Control for Packet Switched Services" explains the load
control mechanism, that is adopted to correctly distribute the relevant
GPRS/EGPRS traffic among the internal resources of the BSC;
• The Chapter:"9 GPRS/EGPRS Procedures"describes the main procedures
regarding the packet switched services (PS), such as:
– GPRS/EGPRS attach/detach.
– PDP context activation/deactivation.
– Packet Data transfer main procedures.
• The Chapter: "10 GPRS/EGPRS Functionalities" introduces the more recent
GPRS/EGPRS algorithms regarding:
– cell selection/re-selection;
– traffic control management;
– power control;
– link adaptation;
• The Chapter:"11 Database Parameters and Objects" is like an attachment because
it contains tables that collect all the managed objects and related parameters
discussed in the manual.

14 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

2 Siemens Features Description


This chapter is the collection of all the features related to the GPRS/EGPRS analyzed
and implemented up to the current SBS BR 7.0 release. Each feature is related to its
Feature Sheets (FSH) or corrresponding Change Request (CR). Besides for each
Feature Sheet (or Change Request) the following additional informations are provided:
– FSH number and title.
– a summary of the content.
– the release in which the feature has been implemented.

2.1 BR5.5 Feature Description


FSH 0720

GPRS: HW and Basic SW for Packet Control Unit (PCU)


Release BR5.5
This feature is the most important one for the GPRS technology; it describes the
objects, the parameters and the functionalities regarding the packet switched data
service (PS).

CR - F017

Packet Downlink Assignment Procedure on CCCH


Release BR5.5
This Change Request introduces the Packet Downlink Assignment procedure, which is
mandatory, on the CCCH channel.

CR - F135

GPRS Alignment to SMG 30, 30BIS, 31, 31BIS


Release BR5.5
This Change Request aligns the system to the last version of the ETSI standard for the
release ‘97.

CR - F187

GPRS: Non signalling Channels PDCH Static Allocation


Release: BR5.5
This Change Request introduces the possibility to configure static GPRS channels (not
PBCCH or PCCCH) to support data traffic only.

CR - F189

GPRS Improvements Step 1

A30808-X3247-L24-5-7618 15
GPRS/EGPRS Global Description Information
System

Release: BR5.5
This Change Request introduces some improvements regarding the GPRS service,
with the purpose to increase mainly the customer acceptance and performance of the
GPRS.

CR - F190

Support of Non-DRX Mode after Change to Packet Idle Mode


Release: BR5.5
This Change Request allows the reduction of about 50% of the time that is needed to
send data blocks from the Gb interface to the Mobile Station. The target is reached by
accelerating the packet downlink assignment procedure.

CR - F191

Improve Robustness of GPRS Packet DL Assignments


Release: BR5.5
This Change Request allows the reduction of the delay that occurs between the trans-
mission of downlink assignment messages and the beginning of packet downlink data
transfers (in a first step, see CR - F190, the delay that characterized downlink assign-
ment procedures has been reduced. However the delay can still be on average
reduced by 50% with the realization of this CR).

CR - F205

GPRS Improvements Step 2


Release: BR5.5
This Change Request introduces some improvements regarding GPRS service, to
increase customer acceptance and performance of GPRS.

CR - F287

Decrease Round Trip Delay Time and Improve Web Browsing Performances
Release BR5.5
This Change Request allows the improvement in the overall performance of the inter-
action between many TCP/IP based applications and the GPRS network.

CR - X232

GPRS Improvements for BR5.5


Release: BR5.5
This Change Request allows the improvement in the GPRS network, by introducing the
following features without O&M impacts:
- GPRS channels on all the TRXs of a cell;
- horizontal allocation.

16 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

CR - X366

Change Polling Strategy during Delay TBF Release


Release: BR5.5
This Change Request allows the reduction the time needed by the MS to establish a
concurrent uplink TBF, when the downlink TBF is kept open using the Delay TBF
release procedure, introduced by CR - F287.

2.2 BR6.0 Feature Description


FSH 0397

High Capacity BSC


Release BR6.0
This feature introduces the first step for the High Capacity BSC (HC BSC step1), that
exploits the rack already used in the previous releases.

FSH 0457

Service Dependent Channel Allocation Strategy - Step1


Release BR6.0
This feature introduces new strategies to manage Circuit Switched, GPRS and HSCSD
calls.

FSH 0503

GPRS: Automatic Horizontal Allocation


Release BR6.0
This feature introduces the horizontal allocation strategy, and the parameter used to
handle it.

FSH 0512

Packet Transfer on non BCCH TRXs without Downlink Power Control


Release BR6.0
This feature introduces the possibility to configure the GPRS service on the TRXs
chosen by the operator.

A30808-X3247-L24-5-7618 17
GPRS/EGPRS Global Description Information
System

FSH 0515

Improvement in GPRS scheduler


Release BR6.0
This feature introduces a new mechanism to schedule data blocks to be sent/received
to/from the users.

FSH 1928

Miscellaneous Impacts from Q3IG and DIG


Release BR6.0
This feature introduces the new method to configure both intra-BSC and inter-BSC
neighboring cells.

CR - F092

Implementation of FRS0457: Service Dependent Channel Allocation Strategy -


Step 1
Release: BR6.0
This Change Request allows implementation of the Service Dependent Channel Allo-
cation Strategy - Step 1, described in FSH 0457, in BR6.0 release.

CR - F119

Update of FRS 1928 (Miscellaneous impacts from Q3IG and DIG)


Release: BR6.0
This Change Request is an update of FSH 1928.

CR - F208

Rework of default values for Power Control, Handover, Adjacent Cell and BTS
Release: BR6.0
This Change Request introduces new default values for some parameters.

CR - X260

GSM-UMTS Cell Selection/Re-Selection


Release: BR6.0
This Change Request allows GSM/UMTS users to perform a cell reselection from GSM
cells to UMTS cells.

CR - X263

GPRS scheduler Modification


Release: BR6.0

18 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

This Change Request is due to the decision to implement only a few parts of FSH 0515.

CR - X411

Removal of limitation in the number of GPRS adjacent cells


Release: BR6.0
This Change Request increases to 32 the maximum number of GSM adjacent cells
supporting GPRS.

CR - X482

UMTS-GPRS Cell reselection


Release: BR6.0
This Change Request allows GPRS users to perform a cell reselection from GSM cells
to UMTS cells without loosing the service.

CR - X617

New Attribute Definition and Default Adjustment


Release: BR6.0
This Change Request introduces the TIMEDTBFREL parameter and new default
values for some GPRS parameters.

CR - X669

GPRS Resume Procedure


Release: BR6.0
This Change Request allows implementation of the GPRS resume procedure already
foreseen for next releases.

CR - X685

New PTPPKF Object Management in Case of Unavailability of TRXs Supporting


GPRS
Release: BR6.0
This Change Request establishes that when all the TRXs supporting GPRS in a cell
are excluded from the service, because LOCKED and/or disabled, the related PTPPKF
object instance is excluded from service too (put into DISABLED state).

CR - X706

Reserved GPRS Channels Management Modification


Release: BR6.0
This Change Request introduces the GMAPERTCHRES parameter and a new defini-
tion for the GDCH one.

A30808-X3247-L24-5-7618 19
GPRS/EGPRS Global Description Information
System

CR - X912

Modification of “Busy Traffic Channel” Calculation


Release: BR6.0
This Change Request establishes that GPRS non reserved channels must be taken
into account in the calculation of BUSY TRAFFIC CHANNELS only if the setting of the
DGRSTRGY parameter does not allow GPRS downgrade.

CR - X1086

GPRS - Uplink Balanced Assignment of Resources


Release: BR6.0
This Change Request allows to implement an Uplink Balanced assignment of
resources; so it will be possible to vary the number of timeslots assigned in Uplink direc-
tion on TBF basis for those MSs that support the dynamic allocation of resources in
downlink and uplink directions (Multi-slot mobiles class 6,7,10,11 and 12).

CR - X1519

Enable Throughput of GPRS Attach/Detach Requests to/from Rel. 99


Release: BR6.0
This Change Request allows stopping discarding the ATTACH ACCEPT message
when it contains optional fields; without this CR the discarding happened when using
Rel. 99 Handset in BR6.0 (Rel. 98) networks.

CR - X1553

Network Adaptation to Present MS Implementation Regarding PCCCH Operation


Release: BR6.0
This Change Request assures that PPCH is not used for data transfer in CS-2; other-
wise the MS will not be able to decode the radio blocks, and interprets them as
"corrupted". Then the downlink signalling counter expires, causing reselection to
another cell or loss of connection.

CR - X1681

Enlarge Fast Polling Period During Delayed DL TBF


Release: BR6.0
This Change Request allows speeding up the MS uplink establishment during the
Delay DL TBF Release Time. Improvements of about 80 ms are expected on Ping
Delay time and also on FTP throughput.

20 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

CR - X1706

Possibility to Enable/Disable Peak Throughput Management Feature


Release: BR6.0
This Change Request allows the possibility to Operator to enable/disable the Peak
Throughput Management feature. This permits to reduce the time to perform the GPRS
Attach procedure.

2.3 BR7.0 Feature Description


FSH 0418

GPRS: Network Controlled Cell Reselection


Release BR7.0
This feature introduces new strategiesfor the management of both GPRS and EGPRS
packet switched data traffic. It introduces new parameters for packet switched cell re-
selection from GSM to UMTS.

FSH 0419

Support of CS3, CS4


Release BR7.0
This feature introduces new GPRS coding schemes.

FSH 0420

MAC Protocol Enhancements for EDGE


Release BR7.0
This feature introduces the enhancements regarding the MAC protocol that is used for
the support of the EDGE functionality. The feature sheet also comprises enhancements
regarding both RLC and BSSGP protocols.
It introduces some EDGE parameters related to the previous protocols and new flags
to enable and disable GPRS and EGPRS on a cell basis.

FSH 0429

EDGE: Flexible Abis Allocation Strategy (FAAS)


Release BR7.0
This feature introduces a new strategy to manage resources of the Abis interface. With
this strategy it is possible to assign in a dynamic way, more than one Abis subslot to a
single air timeslot. New PCU frames are also defined.

A30808-X3247-L24-5-7618 21
GPRS/EGPRS Global Description Information
System

FSH 0444

Link Quality Control: (LA)


Release BR7.0
This feature introduces the link adaptation algorithms regarding both GPRS and
EGPRS services.

FSH 0514

Gb/MS flow control (GPRS Step 1 Completion)


Release BR7.0
This feature describes the flow control procedure on the Gb interface.

FSH 0516

GPRS Resource Management


Release BR7.0
This feature introduces new algorithms to manage radio resources when packet
switched services (GPRS/EGPRS) are enabled.

FSH 0527

2nd Step of the High Capacity BSC


Release BR7.0
This feature introduces a further enhancement of the High Capacity BSC step1 imple-
mented in BR6.0, based on new rack and boards.It’s called: “2nd step of the High
Capacity BSC”.

FSH 0550

EGPRS/GPRS Scheduler Enhancements


Release BR7.0
This feature introduces enhancements in the process that manages the transmis-
sion/reception of GPRS/EGPRS radio blocks on the radio interface.

CR - X0158

Enable/Disable GPRS and/or EDGE Support on Call Basis


Release: BR7.0
This Change Request introduces two new attributes related to the PTPPKF object, to
enable GPRS and EGPRS services on cell basis.

22 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

CR - X1150

Improvement of CS Channel Allocation


Release: BR7.0
This Change Request introduces changes in PDCH pre-emption, when a circuit
switched call must be served in a congested cell.

CR - X1152

Adaptation of FRS AEK0514A to the Current Implementation


Release: BR7.0
The current description of FRS AEK514 (MS/Gb flow control) does not reflect the
current implementation. It contains the general and unuseful information that some
parameters are under investigation. The requirement contained in this Change
Request asks the alignement of the FRS to the current implementation (Feature
Sheet).

CR - X1362

New O&M Attributes for Network Controlled Cell Reselection


Release: BR7.0
This Change Request asks the addition of the following attributes that are necessary
for the implementation of the feature: Network Controlled Cell Reselection
ADJC object:
NCGRESOFF,NCGTEMPOFF,NCGPENTIME,NCC1THRSADJC
PTPPKF object:
NCC1THRS,NCSAMERA,NCRARESH .
The following attribute have to be added for the Handover from GSM to UMTS
BSC object:
BSCT3121

CR - X1454

Multiplexing of GPRS and EGPRS on the same Timeslot


Release: BR7.0
This Change Request allows multiplexing of GPRS and EGPRS mobile stations on the
same PDCH dynamically.The possibility of multiplexing GPRS and EGPRS mobiles on
the same channel enables that the customer is not forced to have separeted channels
for those mobile types. Especially Cingular and Vodafone D2 is asking for this function-
ality.Without this feature Cingular fears, that they cannot meet their business case due
to waste of resources.

A30808-X3247-L24-5-7618 23
GPRS/EGPRS Global Description Information
System

CR - X1495

Uplink Balanced Assignment of E(GPRS) Resources


Release: BR7.0
This Change Request introduces a new strategy to better manage concurrent TBFs, for
Mobile Staions able to use more than one timeslot in the uplink direction.Mobile
stations (MS) which provides a dynamic allocation of the number of uplink and downlink
time slots (multislot class 6, 7, 10, 11 and 12) should be able to use the maximum
number of time slots in uplink direction compatible with dynamic allocation for data
transfer, if there is a considerable amount of uplink traffic available. The MS indicates
the amount of uplink data with a special parameter in the channel request description
and the newtwork should take this parameter into account by assigning the time slots
for both uplink and downlink TBFs. Tests with a multislot class 6 MS have shown, that
with two simultaneous ftp connections, one in uplink the other in downlink direction
(duplex FTP), in case of downlink preferred configuration (3+1) the downlink
throughput is worse than in uplink preferred configuration (2+2). This is due to the fact
that ftp connections are based on TCP as transfer protocol, which causes as acknowl-
edged protocol also traffic in the opposite direction. Because of the delayed
acknoledgement packets (caused by the queue in MS or notebook which is always full
concerning the uplink traffic) the downlink transfer is reduced (stalled condition).
Also pure UL traffic like FTP put is not handeled optimally, since the network changes
to downlink preferred allocation as soon as first DL TBFs for TCP/IP acknowledgments
arrive.

CR - X1507

GPRS Improvements on Ping Delay


Release: BR7.0
With this CR it is requested to decrease the Ping Delay Time by reducing the internal
PCU queue from 3 to 1(or max2) radio blocks, that means more or less no internal
queueing but immediate sending of data when available. This improvement can save
approximately 20-40 msec per direction and it's requested only on PPXU.

CR - X1656

Shortening of Duration of Immediate Assignment Procedure for GPRS


Release: BR7.0
This Change Request allows to transmit immediate assignment message after that, on
each PDCH involved in TBF and which has to be aligned, only two complete uplink
PCU frame have been received by BSC. In this way the duration of Immediate Assign-
ment procedure is reduced.

24 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

CR - X1738

GPRS/EGPRS: Improvement of ABIC/PCUX Interface to Shorten PCU reaction


time
Release: BR7.0
With this Change Request it's requested to furtherly shorten PCU reaction times (round
trip Delay on Abis, Abis-GB and Gb-Abis crossing times), by improving AbIC-PCUX
interface.This can be achieved with the following modifications:
1) Delay of Up Link frame interrupt to 4 ms (Currently 10 ms). This means that the UL
frames are segmented into 4 parts, allowing the parallel processing between receiving
information from the Abis and sending data to Pentium (SW modifications required on
both AbIC and PCUX).
2) DownLink (DL) Frame Request from DSP moved to roughly 9 - 6.5 ms before starting
Abis transmission (currently 20 - 16 ms). This allows the parallel processing between
receiving data from Pentium and transmission over ABIS interface (SW modifications
required on AbIC only).
3) DL Block scheduler activated around 5 - 7.5 ms before DL Frame Request (currently
20 - 16 ms). The granularity of DL interrupt is lowered to 2.5 msec, allowing the PCUX
to implement a very precise and timing mechanism.This change also increases the
probability that a GB downlink data is transmitted over the first block requested by DSP
(SW modifications required on both AbIC and PCUX). The coexistence of these
improvements should further save (with respect to BR 7.0 Step 1) approximately 40
msec on PCU round trip time, 20-25 msec on Gb-Abis and some msec on Abis-Gb
crossing time, and it's feasible only on PPXU
Important points to be outlined are the following:
- Timing is referred to a low_load condition;
- The new performance is highly challenging and requires a complex tuning between
PCUX Operating System, AbIC and UPK. With the current PPXU card, no further
improvements is possible.

CR - X1742

Enable/Disable of Delayed DL TBF During Mobility Management

A30808-X3247-L24-5-7618 25
GPRS/EGPRS Global Description Information
System

Release: BR7.0
IIn this Change Request it is requested to maintaine active the delayed TBF also during
Mobility management procedures considering always the peak throughput information
for the assignment of resourses.
The Multislot class will be considered only if peak throughput information has inconsis-
tent values as already planned (e.g.0).
It's requested to give the possibility to the operator to decide if enable or disable
delayed TBF release during Mobility Management procedure in order not to have prob-
lems with other customers. When delayed TBF release during MM procedures is
disabled (current implementation):
the assignment of resourses is done using the peak throughput information every time
the TBF is established for the different procedures (signalling or data). This means for
example that 1 TS will be assigned for signalling procedures and more TS for data
depending on the information sent by SGSN. esourses).
When delayed TBF release during MM procedures is enabled (to optimise GPRS
attach time):
the assignement of resourses is done using the peak throughput information but the
TBF in this case is maintained active during "transaction" from signaling to data. Taking
the example made before, this means that 1 TS will be assigned for signalling proce-
dures and an upgrade procedure will be activated on the same TBF to assign more TS
for data.

CR - X1850

No “ping_pong” behaviour for mobiles which do not transmit packet cell change
failure
Release: BR7.0
This Change Request allows to prevent “ping_pong” effect due to questionable Mobile
Station behaviour during Netowrk Controlled Cell Reselection.To handle this event the
BSC has not to order to mobile to move again into this adjacent target cell , in
spite of good radio link scenario , until the timer TRFPSCTRL is expired .
This action trust in the fact that mobile’s TLLI used in the old serving cell and mobile’s
TLLI used in the adjacent target cell may differ only for one bit ( bit 30th ,which distin-
guish between local / foreign TLLI ) , otherwise BSC may not track mobile in
its cell change.This procedure requires also that BSC stores informations related to
mobile after the end of each TBF at least for the time STGTTLLIINF ( storage TLLI Info).

CR - X1869

Disable CS3&CS4
Release: BR7.0
Siemens is introducing the GPRS CS3&CS4 in BR7.0. Currently the CS3&CS4 featrue
is dependent on the EDGE activation. It is activated when EDGE is on and de-activated
when EDGE is off. The Siemens customer Cingular has decided not to launch GPRS
with CS3&CS4 in all their markets. Therefore this CR allows to enable/disable the
CS3&CS4 feature independently from EDGE feature activation

26 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

CR-X2132

Title . Removal of BSS restrictions in extended band.


Release: BR7.0
Description: This Change Request requires to remove some limitations on
the GPRS and FHSY.For the GPRS the GSUP can be set also
in a different band that the BCCH one. It is left to the user
choiche the decision to enable the GPRS on the Extended
band. For FHSY the new implementation allows the configura-
tion of FH laws in the extended band overlapping into the
primary band. No overlapping between primary/extended band
and DCS shall be kept

CR X-2199

Title: Common BCCH Improvements for 900 and 1800Mhz


Release: BR 7.0
Description: This Change Requests asks the implementation of the patch
solution described in the Change Request 1300 (Common
BCCH for Cingular - extension to BR 7.0) also for the frequen-
cies 900 / 1800. The CR1300 asks to extend to the release BR
7.0 the patch provided with the CR 688: “Modification of
Common BCCH Implementation via Patch”. The customer
asks the use of the common BCCH and (E)GPRS in the inner
area.Besides currently there are in progress commercial nego-
tiations with the customer Eurotel in Czech Republic. Siemens
has the big opportunity to swap NOKIA out of the customer’s
network. For the reason that Nokia has already implemented
the Common BCCH feature”, the patch solution described in
the Change Request is the precondition for a successful
commercial strategy.

CR X-2230

Title: Enhancement of throughput of 8PSK Mobile Stations when


multiplexed with GMSK Mobile Stations
Release: BR 7.0
Description: This change request asks to implement the requirements
described in the FRS AEK550. The throughput decreasing of
EDGE Mobile Station in 8PSK modulation could be at
maximum at 35% with BLER = 0 (no interference) and at
maximum 31% in the normal field condition while currrently its
degrading is near to 70%.

CR X-2263

Title: Common BCCH allowing E(GPRS) in the complementary


Band.
Release: BR 7.0
Description: This Change Request enables the usage of a common BCCH
on the 900MHz frequency and it allows the releases of the
GPRS/EGPRS services on both 900 and 1800 MHz.

A30808-X3247-L24-5-7618 27
GPRS/EGPRS Global Description Information
System

CR X-2313

Title: Enable Directed Retry to UMTS Independent of Enable Imper-


ative Handover HO).
Release: BR 7.0
Description: This Change Request asks to enable/disable the Directed
Retry to UMTS independently from the setting of the
enable/disable flag of the imperative Handover to UMTS. The
FRS AEK0490 shall be updated accordingly to this Change
Request.

CR X-2325

Title: (E) GPRS improvements on first ping and gap between IAMCD
and PRR/TBF start
Release: BR7.0
Description: This Change Request asks the following improvements for the
GPRS/EGPRS system: 1) For the First Ping the number of the
PDT assigned to a single block has to be set to 2 if concate-
nated PCU frames are used in the cell and to 1 if standard PCU
frames are used. 2) In the current load there is a gap of roughly
350-450ms between the IACMD and the PRR in case of 2
phase access. For this reason a reduction/optimization of the
overall delay for all kinds of the PRR/TBF start has to be
applied for cases with both idle channels as weel as active
channels. 3) In the current load the BCCH change mark is
changed about every 15 minutes to refresh the system info.
The Mobile Station will release the ongoing TBF to read all the
incoming systeminfo even if they are not changed. Therefore it
is requested to enlarge the repetition rate to refresh the system
info in order to decrease the number of the TBF released in the
network. 4) Improvement and Optimization of the GPRS and
EGPRS Link Adaptation Thresholds.

28 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

3 GPRS/EGPRS Overview
The General Packet Radio Service (GPRS) and the Enhanced General Packet Radio
Service (EGPRS) allow packet switched data transmission on the framework provided
by the GSM mobile network.
The well known word EDGE (Enhanced Data rates for the GSM Evolution) applies both
i to the circuit switched (CS) and to the packet switched (PS) services. Note that EDGE
is mainly a characteristic of the Air Interface, including a new kind of modulation (8PSK,
besides the already used GMSK modulation. See for more details the chapter:
"3.1 GPRS and EGPRS Modulation Principles").
The word EGPRS (Enhanced GPRS) applies only to the packet switched (PS) services.
Whenever in this document the word EGPRS is adopted, EDGE is referred and it is
applied to the packet switched (PS) services. That means, substantially, the coding of
the radio blocks using a specific set of modulation and coding schemes (MCS1, ..,
MCS9), and using new specific RLC/MAC control messages or new specific information
elements in GPRS RLC/MAC control messages. In the current BR 7.0 release, EDGE
is applied only to packet services. However, the generic term EDGE is used in O&M
attributes that, in some future release, could be used to define the support of EDGE also
for the circuit switched (CS) service. In the remainder of this manual, the world EDGE
means EGPRS and viceversa.

When the GPRS/EGPRS technology is not configured, the GSM/DCS network works in
circuit switched connection mode, i.e., it gives to the customer the exclusive use of a
certain amount of bandwidth for the duration of the requirement. The connection is set
up on demand and released when the caller breaks the connection. Circuit switched
connections (CS) are what is provided by the GSM architecture for speech and data
services. Data transmission with bandwidth larger than 9.6 kbit/s (or larger than 14.4
kbit/s, if this higher data rate is enabled) is reached by combining more radio channels
to a given user, by the HSCSD feature. Nevertheless, when a circuit switched connec-
tion is established and the user does not transmitt information, which is typical of data
transmission, the specific resources are wasted because they are not available for other
users requesting the services. In other words, it means that circuit switched connections
do not provide an optimized way to support data traffic.
In order to improve and optimize the use of both the network and radio resources, for
both GPRS and EGPRS technology the packet switched (PS) technique has been
implemented for supporting both data and signalling transfer in an efficient manner.
New GPRS/EGPRS radio channels are defined, and the allocation of these channels is
flexible as follow:
– from 1 to 8 radio interface timeslots can be allocated for TDMA frame, for each trans-
ceiver of the cell;
– timeslots are shared by the active users (i.e., the same timeslot can be assigned to
different users at the same time, unlike what happens in GSM);
– radio interface resources can be shared dynamically between speech services (i.e.,
circuit switched services) and data services (i.e., packet switched services) as a
function of service load and also on the basis of different operator’s needs;
– uplink and downlink resources are allocated separately.
Applications that take advantage of GPRS/EGPRS services should exhibit one or more
of the following characteristics:
• intermittent, non-periodic (i.e., bursts) data transmission;
• frequent transmission of small volumes of data;
• not frequent transmission of large volumes of data.

A30808-X3247-L24-5-7618 29
GPRS/EGPRS Global Description Information
System

3.1 GPRS and EGPRS Modulation Principles


The GPRS technology is an evolution of the existing GSM technology and it uses the
same modulation scheme, called GMSK (Gaussian Minimum Shift Keying). The GMSK
digital modulation format relies on shifting the carrier 180˚ in phase to produce a binary
modulation scheme capable of delivering 1 bit/symbol (see Fig. 3.1).

Fig. 3.1 Basic GMSK Constellation of Signal Vectors

The GPRS uses four different channel coding schemes (see the chapter: "4.2.1 GPRS
Channel Coding") to provide different levels of protection to the packets on the air inter-
face.
This modulation scheme, within 200 KHz bandwidth, provides good spectral perfor-
mance and an adequate data rates for GSM voice applications, however it cannot
supply fast data services since it only transmits 1 bit/symbol.
The EDGE technology uses the same bandwidth allocated for GSM voice and GPRS
data services, but delivers a higher capacity and fast data services to the mobile network
by using a new modulation scheme called 8 PSK (8-level Phase Shift Keying). With this
8PSK modulation, there are eight distinct phase changes that the decoder will look for
the conversion into binary data. Each phase represents a symbol and carries three bits
of information. (see Fig. 3.2).

30 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Fig. 3.2 Basic 8 PSK Constellation of Signal Vectors

As a consequence, the EDGE’s 8 level-shift keying modulation scheme allows a radio


throughput increase of almost 3 times the radio throughput of GPRS with the same
number of timeslots with big advantages for the final users. In the following table a
comparison of the physical layer parameters is depicted.

GSM EDGE
Modulation GMSK, 1bit/sym 8 PSK, 3 bit/sym
Symbol Rate 270833 kbit/s 270833 kbit/s
Payload per Burst 114 bit 348 bit
Gross Rate per Time 22.8 kbit/s 69.6 kbit/s
Slot
With the classical 8 PSK modulation scheme, it is possible during symbol changes for
the signal trajectory to pass through the origin (I/Q value 0,0), which causes both a very
high Peak to Average Value (PTA) and a high dynamic range of the signal. To avoid this
possibility, EDGE uses a 3pi/8-shifted 8PSK approach, by which with every phase tran-
sition, the symbols rotate by 3pi/8 causing a shift of the I/Q constellation relative to its
previous starting position.
Nine coding schemes (from MCS1 to MCS9, as described in the chapter: "4.2.2 EGPRS
Channel Coding") using both GMSK and 8PSK modulations are introduced and a link
adaptation algorithm allows automatic switching between coding schemes, based on
the radio environment condition. The Tab. 3.1 shows which EDGE coding schemes are
GMSK modulated and which are 8 PSK modulated.

GMSK modulated 8 PSK modulated

MCS1 MCS5
MCS2 MCS6
MCS3 MCS7

Tab. 3.1 EGPRS Coding Schemes and their Modulation

A30808-X3247-L24-5-7618 31
GPRS/EGPRS Global Description Information
System

GMSK modulated 8 PSK modulated

MCS4 MCS8
MCS9

Tab. 3.1 EGPRS Coding Schemes and their Modulation

3.2 Network Architecture


Generally the packet data network establishes a logical connection between the users
but does not guarantee an immediate access to the transmission network: when more
users ask the access to the transmission resources at the same time, the network has
to schedule the access keeping some of them in a wait queue for avoiding traffic conges-
tion.
As shown in Fig. 3.3, the GPRS/EGPRS network is put on the top of the GSM existing
one but without substitute it. In fact the network architecture still grants that speech and
data transmission with circuit switched connections (CS) are controlled by the MSC
(through the A interface).

Fig. 3.3 GPRS/EGPRS Network Architecture.

But for providing the Packet Switched (PS) services two new network nodes in the GSM
core network have to be introduced:
• Serving GPRS Support Node (SGSN): the SGSN keeps track of the individual
Mobile Station location and performs security functions and access control. It is at
the same hierarchical level as the MSC and it can be connected to the Base Station
System (BSS) via a Frame Relay network. It is also possible to connect the SGSN

32 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

and the BSS via nailed-up connections (NUCs) or through point-to-point connec-
tions.
The SISGSNREL99 parameter is broadcasted in the cell, in order to inform
i the Mobile Stations about the specification Release implementation in the
SGSN.

The SGSN and the BSC are connected through the Gb interface. It specifies the
data flow and the requested protocols (see the Chapter: "7 Gb Interface") and
consists of connections which carry both data and signalling simultaneously, using
the Frame Relay protocol. Besides the Gb interface is “standard” and it guarantees
multi-vendor capabilities.
• Gateway GPRS Support Node (GGSN): the new node GGSN provides:
– interworking with external packet switched (PS) networks;
– management of IP addresses.
The GGSN could be connected to the SGSN via an IP-based GPRS/EGPRS back-
bone network, but these two entities can also reside on the same physical node.
The interface between the SGSN and the GGSN is the “Gn” Interface.Two GGSN nodes
can be interconnected through the so-called “Gp” Interface.
Besides the HLR has to be upgraded with GPRS/EGPRS subscriber information, and
optionally the MSC/VLR can be enhanced for a more efficient coordination of GPRS and
non-GPRS services and functionalities like for example the following:
– paging of circuit switched calls through the SGSN;
– combined GPRS and non-GPRS location updates.
To allow co-ordination of activities between the MSC and the SGSN, the Gs interface
must be supported (see Fig. 3.3).
The security management functions for the GPRS/EGPRS technology do not differ for
those implemented for the GSM system: the SGSN performs authentication and cipher
setting procedures based on the same algorithms, keys, and criteria adopted in GSM;
the only difference is that GPRS/EGPRS networks require a ciphering algorithm opti-
mized for packet data transmission.
In order to access to the packet switched (PS) services, a Mobile Station (a specifc hard-
ware and software is needed for being able to provide GPRS services) first makes its
presence known to the SGSN by performing a GPRS attach procedure. It is described
in detail in the chapter: "9.3.2.1 Attach Function".This operation establishes a logical link
between the Mobile Station and the SGSN, and it provides the following functions:
– paging via the SGSN;
– notification of incoming GPRS/EGPRS specific data;
– SMS over GPRS;
So at the end of a successful GPRS attach procedure, the SGSN establishes with the
mobile station a mobility management session, containing information pertaining to, for
example, mobility and security etc.
In order to send and receive packet switched (PS) data, the Mobile Station first activates
the packet data address that it wants to use. In this way the Mobile Station will be recog-
nized by the corresponding GGSN and then interworking with external data networks
can begin. During this procedure, which is called PDP context activation (i.e., Packet
Data Protocol context activation), the SGSN establishes a PDP context with the related
GGSN as it is described in the chapter: "9.7 Activation and Deactivation of a PDP
Context" This context is used for routing purposes when the user:
– will send data to the external data network;

A30808-X3247-L24-5-7618 33
GPRS/EGPRS Global Description Information
System

– will receive data from the external data network.


At the end of the successful execution of the attach and of the PDP context activation
procedures, the MS can start the transmission or reception of data.
For the purpose, the Mobile Station must establish a physical connection with the
network; this physical connection is called “Temporary Block Flow”. The Temporary
Block Flow allows unidirectional transfer of data through the allocated radio resources.
See for more details the chapter: "4.1 GPRS/EGPRS Physical Channels".
The User data is transferred transparently between the Mobile Station and the external
data networks with a method known as encapsulation and tunnelling: data packets are
completed with GPRS/EGPRS specific protocol information and transferred between
the Mobile Statopm and the GGSN of competence. This transparent transfer method
lessens the requirement for the GPRS PLMN to interpret external data protocols, and it
enables an easy introduction of additional interworking protocols in the future. User data
can be compressed and protected with retransmission protocols, to get a consistent,
efficient and reliable data transmission.

34 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

3.3 GPRS/EGPRS Protocol Stack


The GPRS and EGPRS technology is supported at every level of the OSI stack by a set
of protocols that are represented below (See the figure 3.4: Protocol Stack for Data
Transmission in GPRS/EGPRS Network.) toghether with the corresponding interfaces
starting from the air-interface (“Um”) up to the core Network (“Gn” Interface between the
SGSN and the GGSN).

Fig. 3.4 Protocol Stack for Data Transmission in GPRS/EGPRS Network.

The different layers for the Um, Abis, Gb, Gn and Gi interfaces provide the following
functions:
• GSM RF: the GSM RF is the protocol specified for the Um and the Abis interfaces.
It supports the physical radio channel used to transfer packet data;
• MAC: the Media Access Control layer s the protocol specified for the Um and the
Abis interfaces.It provides the access to the physical radio resources. It is respon-
sible for the physical allocation of the packet data channels (PDCHs);
• RLC: the Radio Link Control layer is the protocol specified for the Um and the Abis
interfaces.It provides a reliable link over the air interface that fits the block structure
of the physical channel; therefore its main task is the segmentation and reassem-
bling of the LLC frames transmitted between the BSS and the SGSN. In addition it
performs:
– a sub-multiplexing to support more than one Mobile Station by one physical
channel;
– the channel combining to provide up to eight physical channels to one Mobile
station.
• LLC: the Logical Link Control layer provides a logical connection between the Mobile
Station and the SGSN even if no physical connection is established. The physical
connection is set up by the RLC/MAC layer when there is data to transmit;

A30808-X3247-L24-5-7618 35
GPRS/EGPRS Global Description Information
System

• BSSGP: the BSSGP protocol is specified for the Gb interface and it is used to
transfer LLC frames together with related information between the SGSN and the
BSC. Such information include QoS (Quality of Service) and routing information;
• SNDCF: the Sub Network Dependent Convergence Protocol is the protocol speci-
fied for the logical interface between the Mobile Station and the SGSN. It performs
the following tasks:
– encryption;
– compression;
– segmentation/re-assembling;
– multiplexing/de-multiplexing of signalling information and data packets.
The encryption function grants the best security for the data transmission whereas
the compression and the segmentation are performed to limit the amount of data
transferred by the LLC layer.
• GTP: The GPRS Tunnelling Protocol is specified for the Gn interface. Its main task
is the encapsulation/de-encapsulation function. The different kinds of data packets
are encapsulated in IP packets since IP is the GPRS/EGPRS internal network
protocol. The encapsulated data packets are then transferred between the GSN
nodes.
• IP/X.25: The network layer represents the network protocol that supports the infor-
mation transferred over the GPRS/EGPRS network starting from the Mobile Station
up to the GGSN. Depending on the supported network protocol (IP, X.25, CLNP),
there are several kinds of network layers;
• Application: The higher layers (for example the “Application Layer”) are outside the
scope of the GPRS/EGPRS, because they are not dependent from the underlying
network.

3.4 Data Flow


This chapter describes the way data is transmitted from the core network (SGSN) up to
the Mobile Station, and vice versa.
The figure 3.5: “RLC/MAC Block and Radio Block Structures” and the figure 3.6: “Data
Flow across Protocol Layers in case of EGPRS(MSC7...MSC9)” represent in which way
the different protocol’s layers handle the data flow:
– The Fig. 3.5 represents the data flow in case of GPRS and EGPRS when the
MSC1..MSC6 coding schemes are used;
– The Fig. 3.6 represents the data flow in case of EGPRS when the MSC7..MSC9
coding schemes are used (see description below);

36 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Fig. 3.5 Data Flow across Protocol Layers in case of GPRS/


EGPRS(MSC1...MSC6)

Fig. 3.6 Data Flow across Protocol Layers in case of EGPRS(MSC7...MSC9)

It is supposed that an IP data packet has to be sent from an external data network to a
mobile subscriber.
Precondition is that the Mobile Station has already executed the “attach” procedure and
i it has already activated the PDP context towards the involved data network.

The following steps are performed:

A30808-X3247-L24-5-7618 37
GPRS/EGPRS Global Description Information
System

1. the Internet Service provider sends the IP data packet unit to the GPRS/EGPRS
network, using the IP address which has been assigned to the Mobile Station during
the PDP context activation procedure;
2. the GGSN searches for the relevant PDP context and forwards the data unit towards
the right SGSN. The original IP data unit is encapsulated in a new one (using the
GTP protocol), and the new IP address is the IP address of the SGSN;
3. the SGSN decapsulates the IP data packet and (by means of the SNDCP protocol)
it subdivides the data packet in a certain number of LLC frames (data is also
encrypted and compressed).
4. when the SGSN knows the location of the Mobile Station (i.e., the cell where the
Mobile Station is camped on), these LLC frames are sent to the right BSC, across
the Gb interface. As in the GSM system, the paging procedure is used to localize the
subscriber.
5. The LLC frames have a variable length; since they have to be sent on the radio inter-
face, which has a limited capacity, the LLC frames are segmented in a certain
number of RLC/MAC blocks; these blocks have a well defined length (according to
the used coding scheme);
6. The RLC/MAC blocks are then sent through the Abis interface, to the right BTS;
RLC/MAC blocks are sent across the Abis interface, by means of PCU frames. Two
i kinds of PCU frames exists:
- standard PCU frames: they allow the transmission of a restricted number of bits every
20 msec and so they support only CS1 and CS2 GPRS coding schemes;
- concatenated PCU frames: they support not only CS1 and CS2 GPRS coding
schemes, but also CS3 and CS4, and all the EGPRS coding schemes (MSC1..MSC9).
More details are described in the chapter: "6.3 PCU Frames and Dynamic Allocation on
the Abis Interface".

7. the BTS executes the following operations for the received RLC/MAC blocks:
– block coding;
– convolutional coding;
– puncturing;
– interleaving.
Regarding these operations, it is important to make a distinction among the following
different cases:
– when GPRS coding schemes are used, a single RLC/MAC block contains one
Information Field only; the BTS executes the described operations on it; after
these operations, each received RLC/MAC block reaches, independently from the
applied coding scheme, a fixed length of 456 bits;
– when EGPRS GMSK coding schemes are used (i.e., from MCS1 to MCS4), a
single contains one Information Field only; the BTS executes the described oper-
ations on it; after these operations, each received RLC/MAC block reaches, inde-
pendently from the applied coding scheme, a fixed length of 1368 bits;
– when EGPRS MCS5 and MCS6 coding schemes are used, a single RLC/MAC
block contains one Information Field only; the BTS executes the described oper-
ations on it; after these operations, each received RLC/MAC block reaches, inde-
pendently from the applied coding scheme, a fixed length of 1392 bits;
– when EGPRS MCS7, MCS8 and MCS9 coding schemes are used, a single
RLC/MAC block contains two Information Fields; the BTS executes the described
operations on the RLC/MAC block; after these operations, the RLC/MAC block

38 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

reaches, independently from the applied coding scheme, a fixed length of 1392
bits;
8. The block that is obtained after different coding procedures is called Radio Block.
Each Radio Block is then sent on the radio interface by means of 4 Normal Bursts,
in fact each Normal Burst can transmit:
– up to 114 bits in cases of GPRS;
– up to 114 bits in cases of EGPRS when GMSK modulation is used;
– up to 348 bits in cases of EGPRS when 8PSK modulation is used.
The figure 3.7: “Data Flow from the SGSN to the MS.” shows the data flow between the
SGSN and the Mobile Station in the downlink direction through the Gb, Abis and Um
interfaces (in the uplink direction the same data flow is transmitted but in the opposite
order).

Fig. 3.7 Data Flow from the SGSN to the MS.

To avoid mis-understanding in this manual, the following definitions are used:


• RLC/MAC block: a RLC/MAC block is a block generated in the BSC (by the
RLC/MAC layer) starting from the LLC-PDU; then this block is sent using PCU
frames towards the BTS that will apply the right coding;
• Radio Block: a Radio Block is a RLC/MAC block that is generated after the BTS has
applied the block coding (i.e., it is a RLC/MAC block plus some coding bits).
After block coding, the BTS will apply the convolutional coding and both puncturing and
i interleaving procedures; after these operations the interested block will reach a fixed
length of 456 or 1392 bits, and it is still called Radio Block.

A30808-X3247-L24-5-7618 39
GPRS/EGPRS Global Description Information
System

3.5 RLC/MAC Block and Radio Block Structures


Different RLC/MAC block (and as a consequence different Radio Block) structures for
data transfer and control message transfer purposes are defined.
The RLC/MAC block structure for data transfer is different between GPRS and EGPRS,
whereas the same RLC/MAC Block structure is used for the management of control
messages.
All the different RLC/MAC block types, after the coding, are always carried by four
i Normal Bursts on the “Um” radio interface.

3.5.1 RLC/MAC and Radio Block Structures: Data Transfer


As it has been described, two different RLC/MAC Block structures are defined for GPRS
and EGPRS data transfer.

3.5.1.1 RLC/MAC Block and Radio Block Structures for GPRS Data Transfer
A RLC/MAC block for data transfer supported by the GPRS technology consists of one
MAC Header, one RLC Header and one RLC Data Block as represented in the
"Fig. 3.8 RLC/MAC block’s structure for Data Transfer".
– The MAC Header contains control fields with different values for the uplink and
downlink directions and it has a constant length of 8 bits.
– The RLC Header contains control fields with different values for the uplink and down-
link directions and it has a variable length;
– the RLC Data Block field contains octets from one or more LLC PDUs.

MAC Header RLC Header RLC Data

Fig. 3.8 RLC/MAC block’s structure for Data Transfer

The RLC/MAC block is sent to the BTS, that will apply a block coding for the error detec-
tion, adding to the RLC Data Block field the “Block Check Sequence (BCS)” field. At the
end of the operation the Radio Block is generated, as represented in the "Fig. 3.9 Radio
Block structure for Data Transfer on the “Um” Interface". This Radio Block, after convo-
lutional coding, puncturing and interleaving, is then transmitted on the “Um” air interface
and carried by four Normal Bursts.

MAC Header RLC Header RLC Data BCS

Fig. 3.9 Radio Block structure for Data Transfer on the “Um” Interface

40 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

3.5.1.2 RLC/MAC Block and Radio Block Structure for EGPRS Data
Transfer
A RLC/MAC block for data transfer supported by the EGPRS technology consists of one
RLC/MAC Header, and one or two RLC Data Blocks.
– the RLC/MAC Header contains control fields with different values for the uplink and
downlink directions. It also has a variable length;
– the RLC Data Block field contains octets from one or more LLC PDUs;The EGPRS
coding schemes from MCS1 to MCS6 use a RLC/MAC block constituted by only one
RLC Data Block field only (as represented in the "Fig. 3.10 RLC/MAC Block struc-
ture for Data Transfer with one RLC Data Block field"), whereas the coding schemes
from MCS7 to MCS9 use a RLC/MAC block constituted by two RLC Data Block fields
to reach a more high data rate as represented in the "Fig. 3.11 RLC/MAC Block
structure for Data Transfer with two RLC Data block fields".

RLC/MAC RLC Data Block


Header

Fig. 3.10 RLC/MAC Block structure for Data Transfer with one RLC Data Block field

RLC/MAC RLC Data Block RLC Data Block


Header

Fig. 3.11 RLC/MAC Block structure for Data Transfer with two RLC Data block fields

The RLC/MAC block is sent to the BTS, that will apply a block coding for the error detec-
tion. At the end of the operation the Radio Block is generated. (see the "Fig. 3.12 Radio
Block for Data Transfer with one RLC Data Block field"in case only one RLC Data Block
is inserted and the "Fig. 3.13 Radio Block for Data Transfer with two RLC Data Block
field" in case two RLC Data Blocks are inserted). Besides two different block codings
are applied for the error detection:
– the Block Check Sequence (BCS) is used for the error detection of the data part.
– the Header Check Sequence (HCS) is used for the error detection of the header
part.
The RLC/MAC Header does not interact from the RLC Data Block and it has its own
check sequence.
In cases of RLC/MAC blocks constituted by two RLC Data Block fields , each field has
its own block check sequence whereas the RLC/MAC Header is common for both the
fields.
At the end of the checks and after convolutional coding, puncturing and interleaving, the
RLC/MAC Block structure represented in the "Fig. 3.13 Radio Block for Data Transfer
with two RLC Data Block field" is transmitted on the “Um” Air Interface and carried by
four Normal Bursts.

A30808-X3247-L24-5-7618 41
GPRS/EGPRS Global Description Information
System

RLC/MAC
Header HCS RLC Data Block BCS

Fig. 3.12 Radio Block for Data Transfer with one RLC Data Block field

RLC/MAC RLC Data Block


HCS RLC Data Block BCS BCS
Header

Fig. 3.13 Radio Block for Data Transfer with two RLC Data Block field

3.5.2 RLC/MAC Block Structure: Control Signalling


The same RLC/MAC Block for transferring a control message (for example a signalling
message) is supported by the GPRS and the EGPRS technology. It consists of one
MAC header and one RLC/MAC Control Message as represented in the
"Fig. 3.14 RLC/MAC Block Structure for Control Messages". The Header and the
RLC?MAC Control Message have the following structure:
– the MAC Header contains control fields with different values for the uplink and down-
link directions and it has a constant length of 8 bits.
– the RLC/MAC Control Message field contains one RLC/MAC control message;
It is always carried by four normal bursts.

MAC Header RLC/MAC Control Message

Fig. 3.14 RLC/MAC Block Structure for Control Messages

The RLC/MAC block is sent to the BTS that will apply a block coding for the error detec-
tion by the addition of a Block Check Sequence (BCS) field. At the end of the operation
the Radio Block is generated as represented in the "Fig. 3.15 Radio Block for Control
Messages (Signalling).". After convolutional coding, puncturing and interleaving the
Radio Block is then transmitted on the “Um” Air interface and carried by four Normal
Bursts.

MAC Header RLC/MAC Control Message BCS

Fig. 3.15 Radio Block for Control Messages (Signalling).

42 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

The following control messages can be transmitted in the downlink direction within a
RLC/MAC Signalling Block Structure:
– Packet Paging Request: This message is sent by the network to trigger the channel
access by up to four Mobile Stations for a connection’ s establishment.
– Packet Downlink Assignment: This message is sent from the network to assign
resources to the Mobile Station in the downlink direction.
– Packet Uplink Ack/Nack: This message is sent from the network to the Mobile
Station for the acknowledgement of data blocks sent in the uplink direction;
– Packet Power Control/Timing Advance: This message is sent by the network to the
Mobile Station for the reconfiguration of either the “timing advance (TA)” and/or the
power control parameters;
– Packet Access Reject: This message is sent by the network to the Mobile Station to
indicate that the network has rejected its access request.
The following control messages can be transmitted in the uplink direction within a
RLC/MAC Signalling Block Structure:
– Packet Downlink Ack/Nack:This message is sent from the Mobile Station to the
network for the acknowledgement of data blocks sent in the downlink direction.
– Packet Control Acknowledgment: This message is sent from the Mobile Station to
the network for the acknowledge of control blocks sent in the downlink direction;
The Packet Control Acknowledgment message is not formatted as a single RLC/MAC
i block, but as four Access Bursts.

A30808-X3247-L24-5-7618 43
GPRS/EGPRS Global Description Information
System

4 Radio Interface Description


For the configuration of the packet switched data (PS) services in a specific cell, the user
shall create the PTPPKF object (Point To Point Packet Function) related to that cell and
then he/she shall configure properly all the related attributes. The operation can be done
locally from the Local Maintenance Terminal (LMT) or from the Network Management
System (Radio Commander) by means of the command: “Create PTPPKF”. This
command creates an instance of the PTPPKF Managed Object Class (MOC). In the
Containment Tree the PTPPKF Managed Object is hierarchically dependent from the
BTS Managed Object. For each BTS instance (that means for each configured cell) it is
defined only one PTPPKF Managed Object Instance (MOI) subordinated to it. Its default
value is always “0”. The configuration’s operation is permitted if the super-ordinated BTS
and at least one instance (but it is recommended to create always all the instances) of
the NSVC (Network Service Virtual Container) Functional Managed Object have been
previously created. This Managed Object models the functional end-to-end communica-
tion between the BSS and the core network (SGSN).
At the end of the PTPPKF Managed Object successfully creation the cell is allowed to
support Packet Switched (PS) services on the basis of the configuration settings
assigned by the user. These settings are specified in the chapter: "5 Radio Resources
Management".

Functional object Meaning

PTPPKF This Functional Managed Object models the Point to


Point service in a cell. It allows a cell to provide Packed
Switched (PS) data services supported by the
GPRS/EDGE technology. Only one istance (default
value: “0”) of the PTPKF Managed Object can be config-
ured for the superior BTS.

Tab. 4.1 PTPPKF Managed Object.

Once packet switched services have been enabled, the radio resources of the cell can
be assigned to either GPRS/EGPRS packet or circuit switched services, accordingly to
the user’s preferences.
In the GPRS/EGPRS system two types of radio channels have been defined:
1. On-demand radio channels (also called dynamic channels): these channels are
shared between packet switched services and circuit switched services accordingly
to the current requests, but circuit switched services have an higher priority than
GPRS/EGPRS packed swtiched ones.
2. Dedicated radio channels (also called static channels): these channels are perma-
nently assigned to GPRS/EGPRS packet switched services, and they cannot be
used for circuit switched services (even if no GPRS/EGPRS users are exploiting
these channels).

4.1 GPRS/EGPRS Physical Channels


The physical channel (one timeslot of the TDMA frame) assigned to the Packet Data
Services (PS) (either statically or dynamically) is named “Packet Data Channel (PDCH)
as represented within the "Fig. 4.1 Packet Data Channel (PDCH) within a TDMA frame"

44 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

TDMA frame
GPRS/
EGPRS
0 7
PDCH

Fig. 4.1 Packet Data Channel (PDCH) within a TDMA frame

When a timeslot is used for GPRS/EGPRS (that means when the timeslot is a PDCH
one), the multiframe structure for this PDCH consists of 52 TDMA frames structured as
follow:
– 12 blocks (one block is composed by 4 frames and it is represented as Bx, with x=
0..11); each block can convey a RLC/MAC Radio Block containing either data or
signalling as described in the chapter: "3.5 RLC/MAC Block and Radio Block Struc-
tures".
– 2 idle frames represented as “I” and used for measurements.
– 2 frames used for the continuous timing advance update procedure described in the
chapter "4.6 Packet Timing Advance Estimation").

52 TDMA Frame - PDCH Multiframe

B0 B1 B2 T B3 B4 B5 i B6 B7 B8 T B9 B10 B11 i

4 frames 1 frame
- i = Idle frame
- Bx = Radio Block
- T = PTCCH

Fig. 4.2 Multiframe Structure for a PDCH

4.2 Channel Coding


The Blocks B0..B11 sent on the radio interface inside the PDCH multiframe are coded
differently depending on the packet switched (PS) service used (GPRS or EGPRS). In
the following chapters the differences between the two services are described from the
coding process point of view.

4.2.1 GPRS Channel Coding


Four coding schemes: “CS1, CS2, CS3 and CS4” are defined for GPRS RLC/MAC
blocks used during the data transmission.
The "Tab. 4.2 GPRS Coding Schemes" below summarizes the main characteristics of
each coding scheme, referring to the structure of the GPRS RLC/MAC block for data
transfer as represented in the "Fig. 3.8 RLC/MAC block’s structure for Data Transfer".

A30808-X3247-L24-5-7618 45
GPRS/EGPRS Global Description Information
System

Coding Bits of RLC Spare Network Data Rate Bits of Total size of
scheme Data Field bits in RLC/MAC the RLC/MAC
(without RLC Data Header block (bits)
spare bits) Field (including
USF)

CS1 160 0 8 kbit/s (160 bit/20 msec) 24 184


CS2 240 7 12 kbit/s (240 bit/20 msec) 24 271
CS3 288 3 14.4 kbit/s (288 bit/20 msec) 24 315
CS4 400 7 20 kbit/s (400 bit/20 msec) 24 431

Tab. 4.2 GPRS Coding Schemes

According to the coding scheme used, the message (RLC/MAC block), delivered by
means of PCU frames to the encoder of the BTS, has a fixed size of (obviously the same
thing is valid for the message delivered from the BTS to the BSC):
– 184 bits in cases of CS1;
– 271 bits in cases of CS2;
– 315 bits in cases of CS3;
– 431 bits in cases of CS4.
The BTS will then execute the following operations (the coding process, for every coding
scheme, is detailed in the "Fig. 4.3 GPRS Coding Process"):
1. the first step of the coding procedure is to add a Block Check Sequence (BCS) for
the error detection;
2. the second step consists of the USF pre-coding (except for CS1);
3. the third step consists of the addition of four tail bits. Then an half rate convolutional
coding for the error correction is applied (for CS4 there is no coding specific for the
error correction);
4. the fourth step consists of the puncturing operation. It is executed with the purpose
of obtaining the target coding rate.

46 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Modulation
CS1
Block code 40 bit Convolutional Interleaving
USF= 3 bit code (R=1/2)
+ 4 tail bit

181 bit 184 bit 228 bit 456 bit

CS2 / CS3
Mod.
USF Block code16 bit Convolutional Puncturing Interleaving
USF= 3 bit + 4 tail bit code (R=1/2)
pre-coding

268 bit 271 bit 274 bit 294 bit 588 bit 456 bit
312 bit 315 bit 318 bit 338 bit 676 bit 456 bit

Modulation
CS4
USF Block code Interleaving
USF= 3 bit 16 bit
pre-coding

428 bit 431 bit 440 bit 456 bit

Fig. 4.3 GPRS Coding Process


In the first implementation of GPRS, CS1 and CS2 coding schemes have been intro-
duced. Standard PCU frames were designed to carry the necessary signalling and data
information between the BSC and the BTS, and the GPRS capacity on the Abis was
limited to 16 kbit/s.
In fact, with standard PCU frames, only 271 bits of data can be transmitted, every 20
msec, in the PCU frame, on the Abis interface.
Since CS3 and CS4 contain a number of data bits higher than 271 (CS3 uses 315 bits,
whereas CS4 uses 431 bits), it was not possible to use them.
To support CS3 and CS4 coding schemes, concatenated PCU frames are introduced
in the system, and the Abis throughput per radio channel (PDCH) is increased to n X
16 kbit/s, using the flexible Abis allocation strategy, as described in the chapter:
"6.3 PCU Frames and Dynamic Allocation on the Abis Interface".
So, regarding the Abis interface, the information is transmitted using two kinds of PCU
frames:
a) Concatenated PCU frames are used when the support of CS3/CS4 is enabled at
both BSC and at BTS level;
b) Standard PCU frames are used when the support of CS3/CS4 is disabled at the
BSC or at the BTS level.
To get more information about concatenated PCU frames and the flexible Abis alloca-
i tion strategy refer to the chapter:"6.3 PCU Frames and Dynamic Allocation on the Abis
Interface".
The PCU frame format (concatenated or standard) is chosen at the Initial Time Align-
ment phase, and cannot be changed dynamically during data transfer. Therefore, in
order to be able to reach the higher coding schemes (CS3/CS4), when CS3/CS4 are
supported at O&M level with the configuration of the related parameters, the selected
PCU frame format is “Concatenated”, even if the initial coding scheme could be
supported by standard PCU frames.

A30808-X3247-L24-5-7618 47
GPRS/EGPRS Global Description Information
System

As default, the CS-1 and CS-2 coding schemes are enabled in the BSS; the BSC capa-
bility to support CS3/CS4 coding schemes can be enabled/disabled by the user. For the
purpose the CSCH3CSCH4SUP attribute of the BSC Managed Object allows the user
to enable/disable CS-3/CS-4 coding schemes at the BSC level.
The user can then enable/disable the support of CS3/CS4 on a cell basis configuring the
CSCH3CSCH4SUP attribute of the PTPPKF Managed Object..
When enabling the CS-3 /CS-4 coding schemes the precondition is that the bit 25 of the
i MNTBMASK attribute has to be set to FALSE, otherwise (bit 25 of MNTBMASK=TRUE)
the max coding scheme usable is forced to CS2 independently from the
CSCH3CSCH4SUP value set to TRUE.

With bit 25 of MNTBMASK set to 1, then the CSCH3CSCH4SUP attribute becomes a


flag for stating if CONCATENATED PCU frames or STANDARD PCU frames will be
used in the whole PTPPKF, that is, on all the TRX supporting EGPRS or GPRS of the
related PTPPKF Managed Object:
- when CSCH3CSCH4SUP is set to TRUE, CONCATENATED PCU frames are used
- when CSCH3CSCH4SUP is set to FALSE, STANDARD PCU frames are used.
Therefore the check that the CSCH3CSCH4SUP attribute has to be set to TRUE for
enabling EGPRS services is kept.

This also means that in a PTPPKFManaged Object so configured:


- bit 25 of the MNTBMASK attribute = 1;
- CSCH3CSCH4SUP = TRUE;
- EEDGE = TRUE;
then the maximum GPRS coding scheme will be CS-2 and CONCATENATED PCU
frames will be used on all the TRX supporting EGPRS or GPRS.

Instead in a PTPPKF Managed Object so configured:


- bit 25 of the MNTBMASK attribute = 1
- CSCH3CSCH4SUP = TRUE;
- EEDGE = FALSE;
then the maximum GPRS coding scheme will be CS-2 and CONCATENATED PCU
frames will be used on all the TRX supporting EGPRS or GPRS.

The MNTBMASK attribute is related also to the feature: “Common Bcch allowing
GPRS/EGPRS in the complementary band” introduced in BR7.0 by the Change
Request 2263.
By means of the bit24 of the MNTBMASK attribute (plus an object patch) the feature can
be enabled also for GSMDCS.
By means of the bit17 of the MNTBMASK attribute the GSUP can be enabled. As a
consequence the TRXMD can be set to EGPRS also in TRX in the E900 sub-band, both
in EXT900 and in GSMDCS cells with BCCH in P900. This implies to send all the
frequencies (P900 and E900) within the “SystemInfo1 message” causing a limitation on
the number of possible frequencies that could be used in the cell up to 22 independently
from their value. The number could be larger than 22 only if the frequencies are well-
distributed. This limitation is applied not only to the cells allowed to the GPRS service
but to all the EXT900 or the GSMDCS cells in the BSC. (Change Request 2132).
By means of both the bit17 and the bit24 toghether of the MNTBMASK attribute, the
GSUP can be enabled. As a consequence the TRXMD can be set to EGPRS in all the

48 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

TRXs of the cells (both 900 and 1800 cells). This implies to send all the frequencies
(P900, E900 and DCS1800) within the “SystemInfo1 message” causing a limitation on
the number of possible frequencies that could be used in the cell up to 16 independently
from their value. The number could be larger than 16 only if the frequencies are well-
distributed. This limitation is applied not only to the cells allowed to the GPRS service
but to all the GSMDCS cells in the BSC.
The usage of the bit17 stand alone and the usage of the bit24 and the bit17 toghether
of the MNTBMASK attribute implies modification in the content/encoding of the
SystemInfo1 message for all the cells in the BSC and consequently of the MobileAlloca-
tions that have to be transmitted to all the BTSs connected to the BSC. For the purpose
a very complex procedure is needed but it is not implemented in the BSC in the current
release. For this reason and considering also that the usage of the bit17 and the bit24
normally is related to a specific cell planning strategy and not to a punctual demand, it
is strongly recommended not to change the bit17 and the bit24 when the bit17 is in use
, if at least one cell is configured in the BSC. Instead this operation is permitted during
the offline generation/conversion of the database.
The user can also indicate, on a cell basis, which coding scheme has to be used as
preferred for the data transmission, when a new transmission is initiated (whereas
signalling uses always the CS-1 coding scheme as described in the chapter:
"4.4.5 Coding of GPRS/EGPRS Logical Channels". For GPRS the user can set the
preferred initial coding scheme configuring the INICSCH attribute. .
The user defines a value of coding scheme to be used when a data transmission starts
i configuring the INICSCH parameter. This value will be used only when the system does
not have any other information to choose the initial coding scheme (more details are
described in the chapter: "10.5.3 Selection of the Candidate Initial Coding Scheme").

Then the link adaptation algorithm (the algorithm is described in the chapter: "10.5 Link
Adaptation"), if enabled, can change the coding scheme of the TBF according to specific
radio conditions. If the link adaptation is not enabled, the initial coding scheme is the only
one used for the data transmission in the cell.
As it is described in the chapter: "6 Hardware and Software Architecture", in order to
support GPRS TBFs with CS3 or CS4 coding schemes, the requirements are the
following:
• Only High Capacity BSC(s) support the CS3/CS4 coding schemes;
• BTS1, BTS+, E-microBTS and PicoBTS, support the CS3/CS4 coding schemes.
The coding process of a RLC/MAC block, using CS1, is shown in the "Fig. 4.4 Coding
of the RLC/MAC Block using CS-1": the 456 bits obtained after BTS coding are sent
across four Normal Burst, carrying 57X2 bits of information each one.
In order to simplify the decoding, the stealing bits of the block are used to indicate the
actual coding scheme (see for more details the chapter: "4.2.2 EGPRS Channel
Coding").

A30808-X3247-L24-5-7618 49
GPRS/EGPRS Global Description Information
System

PCU

RLC/MAC block

USF TFI Data bits BCS Tail

3 bits 181 bits 40 bits 4 bits

R=1/2 Convolutional Code


Encrypted RLC frame

456 bits

456 bits are split in 4


Normal bursts.
Normal Burst
Training
Tail Encrypted bits Encrypted bits Tail Guard
Sequence period

57 bits 57 bits

Stealing bits

Fig. 4.4 Coding of the RLC/MAC Block using CS-1

4.2.2 EGPRS Channel Coding


As it has been described in the chapter "3.1 GPRS and EGPRS Modulation Principles",
the following nine different modulation and coding schemes: “MCS1..MCS9”, are
defined for the EGPRS RLC/MAC Blocks, both GMSK and 8 PSK modulated.
The Tab. 4.3 summarizes the main characteristics of each coding scheme, referring to
the structure of the EGPRS RLC/MAC block for data transfer (see the
"Fig. 3.10 RLC/MAC Block structure for Data Transfer with one RLC Data Block
field"and the "Fig. 3.11 RLC/MAC Block structure for Data Transfer with two RLC Data
block fields").

Coding Bits of RLC Net Data Rate Bits of FBI+E Total size of
scheme Data Field RLC/MAC fields the RLC/MAC
(without Header DL/UL (bits) block DL/UL
spare bits) (including (bits)
USF)

MCS1 176 8,8 kbit/s (176 bit/20 msec) 31/31 2 209

Tab. 4.3 EGPRS Coding Schemes

50 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Coding Bits of RLC Net Data Rate Bits of FBI+E Total size of
scheme Data Field RLC/MAC fields the RLC/MAC
(without Header DL/UL (bits) block DL/UL
spare bits) (including (bits)
USF)

MCS2 224 11,2 kbit/s (224 bit/20 msec) 31/31 2 257


MCS3 296 14.8 kbit/s (296 bit/20 msec) 31/31 2 329
MCS4 352 17.6 kbit/s (420 bit/20 msec) 31/31 2 385
MCS5 448 22.4 kbit/s (448 bit/20 msec) 28/37 2 478/487
MCS6 592 29.6 kbit/s (592 bit/20 msec) 28/37 2 622/631
MCS7 448+448 44.8 kbit/s (896 bit/20 msec) 40/46 2+2 940/946
MCS8 544+544 54.4 kbit/s (1088 bit/20 msec) 40/46 2+2 1132/1138
MCS9 592+592 59.2 kbit/s (1184 bit/20 msec) 40/46 2+2 1228/1234

Tab. 4.3 EGPRS Coding Schemes

According to the coding scheme used, the message (RLC/MAC block) delivered, by
means of PCU frames, to the encoder embedded in the BTS software has a fixed size
as follow:
– 209 bits in cases of MCS1;
– 257 bits in cases of MCS2;
– 329 bits in cases of MCS3;
– 385 bits in cases of MCS4;
– 478 bits in cases of MCS5 in the downlink direction, and 487 bits in cases of MCS5
in the uplink direction;
– 622 bits in cases of MCS6 in the downlink direction, and 631 bits in cases of MCS5
in the uplink direction;
– 940 bits in cases of MCS7 in the downlink direction, and 946 bits in cases of MCS7
in the uplink direction;
– 1132 bits in cases of MCS8 in the downlink direction, and 1138 bits in cases of
MCS8 in the uplink direction;
– 1228 bits in cases of MCS9 in the downlink direction, and 1234 bits in cases of
MCS8 in the uplink direction.
Obviously the message transmitted from the BTS to the BSC has the sane size for the
different MCSs:
The MCSs are divided into different families:
– A;
– Apadding;
– B;
– C.
Each family has a different basic unit of payload: 37 (and 34) octects for the A and Apad-
ding family, 28 octects for the B family and 22 octets for the C family respectively.
Different code rates within a family are achieved by transmitting a different number of
payload units within one Radio Block. For the families A and B, one, two or four payload
units are transmitted, instead for the family C only one or two payload units are trans-
mitted.

A30808-X3247-L24-5-7618 51
GPRS/EGPRS Global Description Information
System

The Tab. 4.2.4 shows the correspondence between the families and the related coding
schemes, whereas the "Fig. 4.5 EGPRS Coding Schemes and Families" represents the
different relationships among families, coding schemes and possible units of payload.

FAMILY CODING SCHEMES

A MSC-3, MSC-6, MSC-9


A Padding MSC-3, MSC-6, MSC-8
B MSC-2, MSC-5, MSC-7
C MSC-1, MSC-4

Tab. 4.2.4EGPRS Coding Schemes and Families

Fig. 4.5 EGPRS Coding Schemes and Families

When 4 payload units are transmitted (MCS7, MCS8 and MCS9), they are split into two
separate RLC data fields of the same RLC/MAC block (that means with separate
sequence numbers and BCSs, as reported in the "Fig. 3.11 RLC/MAC Block structure
for Data Transfer with two RLC Data block fields").

52 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

This can be clearly seen by comparing the "Fig. 4.6 Interleaving of MCS9 Coded Data
into Two Consecutive Normal Bursts" (family A, MCS9) and the "Fig. 4.7 Interleaving of
MCS6 Coded Data into Four Consecutive Normal Bursts" (family A, MCS6).
After the reception of a RLC/MAC block from the BSC, the BTS executes the following
operations:
To ensure strong header protection, the header part of the Radio Block (i.e., the
i RLC/MAC header) is independently coded from the data part of the Radio Block.

1. the first step of the coding procedure of the data part of the RLC/MAC Block is to
add a Block Check Sequence (BCS, 12bits) for the error detection;
2. the second step consists of the addition of six tail bits (TB);
3. the third step is the activation of a 1/3 rate convolutional coding with constraint
length 7 for error correction;
4. the fourth step is the execution of the puncturing operation for obtaining the target
coding rate. The puncturing operation takes advantage of the different puncturing
schemes Pi (where i = 1..3), which has impact on Incremental Redundancy as Link
Quality Control method; the Pi for each MCS corresponds to different puncturing
schemes achieving the same coding rate;
5. As fifth and last step , the radio block is rectangular interleaved over 4 bursts (see
the Fig. 4.6 and the Fig. 4.7). Hence the block length for each RLC block is:
– 4*114 = 456 bit in cases of GMSK modulation;
– 4*348 = 1392 bit in cases of 8 PSK modulation (including stealing symbols).
For MCS8 and MCS9, only the header is interleaved over 4 normal bursts. The data
i blocks are interleaved over 2 bursts only. The MCS7 header and data are interleaved
over 4 bursts.

The coding and puncturing scheme of a RLC/MAC radio block is clearly outlined in the
RLC/MAC header within the Coding and Puncturing Scheme indicator field (CPS).
Depending on coding scheme, three different header types are defined as follow:
• Header type 1 is used with coding scheme MCS7, MCS8 and MCS9;
• Header type 2 is used with coding scheme MCS5 and MCS6;
• Header type 3 is used with coding scheme MCS1, MCS2, MCS3 and MCS4.
The header type of an incoming EGPRS radio block is indicated with stealing bits of the
Normal Bursts:
– 12 stealing bits are used in cases of MCS1, MCS2, MCS3 and MCS4 coding
schemes;
– 8 stealing bits are used in cases of MCS5, MCS6, MCS7, MCS8 and MCS9 coding
schemes.
As it has been described in the chapter: "4.2.1 GPRS Channel Coding", stealing bits (8
bits) are also used to indicate the coding scheme used in cases of GPRS radio blocks.
Stealing bits coding is represented in the Tab. 4.2.5.

Coding Scheme Stealing Bits

GPRS CS1 none


GPRS CS2 1,1,0,0,1,0,0,0
GPRS CS3 0,0,1,0,0,0,0,1

Tab. 4.2.5Coding of Stealing Bits for GPRS and EGPRS Radio Blocks

A30808-X3247-L24-5-7618 53
GPRS/EGPRS Global Description Information
System

Coding Scheme Stealing Bits


GPRS CS4 0,0,0,1,0,1,1,0
MSC1, MSC2, MSC3, MCS4 0,0,0,1,0,1,1,0,0,0,0,0
MSC5, MSC6 0,0,0,0,0,0,0,0
MSC7, MSC8, MSC9 1,1,1,0,0,1,1,1

Tab. 4.2.5Coding of Stealing Bits for GPRS and EGPRS Radio Blocks

There are eight stealing bits for 8PSK mode which indicate four header formats. There
are twelve stealing bits for GMSK mode which indicate two header formats: the first eight
of the twelve stealing bits indicate CS4 to allow Mobile Stations supporting GPRS
services to decode the header type 3 and to read the USF field of the header (more
details about the meaning of the USF field are described in the chapter: "4.3 Temporary
Block Flow" ).
The USF field has eight states, which are represented by a binary 3 bit field in the MAC
Header. The USF is encoded to twelve symbols similarly to GPRS, (that is 12 bits for
GMSK modes and 36 bits for 8PSK modes). The FBI (Final Block Indicator) bit and the
E (Extension) bit do not require extra protection: they are encoded along with the data
part.

Fig. 4.6 Interleaving of MCS9 Coded Data into Two Consecutive Normal Bursts

54 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Fig. 4.7 Interleaving of MCS6 Coded Data into Four Consecutive Normal Bursts

The user can also configure , on a cell basis, the coding scheme that has to be used as
preferred for the data transmission, when a new transmission is initiated (whereas
signalling always uses the CS1 coding scheme, as described in the chapter:
"4.4.5 Coding of GPRS/EGPRS Logical Channels").
The user can set the preferred initial coding scheme with the following parameters:
• in the uplink direction, as it is described in the chapter "9.1 Mobile Stations for
Packet Switched Services", not all the Mobile Stations that support the EGPRS
services support also the 8PSK modulation, therefore:
– the IMCSULNIR8PSK attribute suggests the MCS to be used in the uplink direc-
tion if the Mobile Station supports the 8 PSK modulation in this direction;
– the IMCSULNIRGMSK attribute suggests the MCS to be used in the uplink direc-
tion if the Mobile Station supports only the GMSK modulation in this direction;
• in the downlink direction all the Mobile Stations supporting EGPRS services are
obliged to support the 8 PSK modulation, so the INIMCSDL attribute suggests the
MCS to be used in the downlink direction for all the EGPRS Mobile Stations.
The user has to set a value of the coding scheme to be used when a data transmission
i starts. This value is adopted only when the system does not know any other information
for choosing the initial coding scheme (see the chapter "10.5.3 Selection of the Candi-
date Initial Coding Scheme").

The link adaptation algorithm, if enabled, can change the coding scheme of the TBF
according to the radio conditions. If the link adaptation algorithm is not enabled, the
initial coding scheme is the only one used for the data transmission in the cell (details
about the coding schemes’ management are described in the chapter "10.5 Link Adap-
tation").
For supporting the EGPRS coding schemes, concatenated PCU frames are used in the

A30808-X3247-L24-5-7618 55
GPRS/EGPRS Global Description Information
System

system, and the Abis throughput per radio channel (PDCH) is increased to nx16 kbit/s,
using the flexible Abis allocation strategy as described in the chapter: "6.3 PCU
Frames and Dynamic Allocation on the Abis Interface".

4.3 Temporary Block Flow


The Temporary Block Flow (TBF) is the physical connection, between the Mobile
Station and the network, used to support the unidirectional transfer of LLC PDUs on
packet data physical channels (PDCHs).
The TBF is characterized by a set of allocated radio resources on one or more PDCHs,
and it comprises the transmission of a number of RLC/MAC blocks carrying one or more
LLC PDUs. A TBF is maintained only for the duration of the data transmission.
The TBF is established:
• in uplink direction to transfer data (or signalling) from the Mobile Station to the
network;
• in downlink direction to transfer data (or signalling) from the network to the Mobile
Station.
A TBF can operate in either GPRS or EGPRS mode; the network sets the TBF mode in
the PACKET UPLINK ASSIGNMENT, PACKET DOWNLINK ASSIGNMENT or IMME-
DIATE ASSIGNMENT message (see "9.8 Access to the Network (Establishment of a
TBF)").
The EGPRS TBF mode is supported by EGPRS capable MSs only, see "9.1 Mobile
i Stations for Packet Switched Services".

For each TBF, the network assigns a Temporary Flow Identity (TFI). The assigned TFI
is unique among simultaneous TBFs in the same direction, i.e.:
– TBFs belonging to the same direction of transmission must have different TFI
values;
– TBFs belonging to different directions of transmission could have the same TFI
value.
The TFI is assigned to a Mobile Station in a resource assignment message that
precedes the transfer of LLC frames (both in the uplink and the downlink directions)
belonging to one TBF. The same TFI is included in every RLC header belonging to a
particular TBF, as well as in the control messages associated to the LLC frame transfer
(e.g., acknowledgements), in order to address the RLC entities.
Each TBF is then identified by the TFI together with:
• the direction (UL or DL) in which the RLC data block is sent, in cases of RLC data
block;
• the direction (UL or DL) in which the RLC/MAC control message is sent and the
message type, in cases of RLC/MAC control message.

4.3.1 Multiplexing MSs on the same PDCH: Downlink Direction


A downlink TBF, associated to a single PDCH, is set up giving to the Mobile Station:
– a PDCH (i.e., a timeslot);
– a TFI (each mobile station has its own TFI value).
All the Mobile Stations which have a downlink TBF on the same PDCH execute the
following steps:
1. they read all the downlink blocks inside the multiframe, and decode the TFI value;

56 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

2. if the TFI is not the one assigned to the Mobile Station, the block is skipped;
3. if the TFI is the one assigned to the Mobile Station, this means that the block belongs
to it and then data is taken.
The "Fig. 4.8 Multiplexing Mobile Station on the same PDCH (Downlink)" represents
the mobile station behavior.

Fig. 4.8 Multiplexing Mobile Station on the same PDCH (Downlink)

4.3.2 Multiplexing MSs on the same PDCH: Uplink Direction


An uplink TBF, associated to a single PDCH, is set up with the purpose to provide to the
Mobile Station:
– a PDCH (i.e., a timeslot);
– a TFI (each mobile station has its own TFI value);
– an Uplink State Flag (USF); the Uplink State Flag (USF) is used (on a PDCH basis)
to allow multiplexing of uplink Radio blocks, from a number of MSs, in the same
PDCH.
The USF is used in Dynamic Access Modes (see "9.8.1 Medium Access Modes"),
and comprises 3 bits at the beginning of each Radio Block, that is sent in the down-
link direction (the 3 bits belong to the MAC header, see "3.5 RLC/MAC Block and
Radio Block Structures"). It enables the coding of 8 different USF states which are
used to multiplex the uplink traffic.
Then all the mobile stations, which have an uplink TBF on the same PDCH, execute the
following steps:
1. they read all the downlink blocks inside the multiframe, and decode the USF value;
2. if the USF is the one assigned to the Mobile Station, then the Mobile Station sends
its uplink data on the next uplink block, or on the next four uplink blocks;
3. if the USF is not the one assigned to the Mobile Station, then the Mobile Station
doesn’t send its uplink data on the next uplink block, or on next four uplink blocks.
The "Fig. 4.9 Multiplexing Mobile Station on the same PDCH (Uplink)" represents the
Mobile Station behavior in cases of uplink transmission on the next uplink block.

A30808-X3247-L24-5-7618 57
GPRS/EGPRS Global Description Information
System

Fig. 4.9 Multiplexing Mobile Station on the same PDCH (Uplink)

On PDCHs (not carrying PCCCH, see "4.4.2 Packet Common Control Channel
i (PCCCH)"), eight USF values are used to reserve the uplink to different Mobile Stations,
by means of the following rule:
- 7 USF values are used for 7 Mobile Stations that have established an uplink TBF;
- one USF value is used to allow, to those Mobile Stations that have established a down-
link TBF, the transmission of control blocks in the uplink direction (e.g., to transmit the
Packet Downlink Acknowledge message).
So when the network wants to permit to one Mobile Stations, that doesn’t have an uplink
TBF, to transmit in uplink direction, it sets the USF field to this reserved value. In this
way, the Mobile Stations that have an uplink TBF do not transmit in the next uplink block
(since they don’t find their USF value), while the network informs the Mobile Stations
with the downlink TBF, that it must transmit in the uplink block that the network has
reserved for it. To inform the Mobile Stations the network uses the RRBP field which is
contained in all the downlink blocks (it is contained in the MAC header of both data and
control blocks, see the chapter "3.5 RLC/MAC Block and Radio Block Structures"); with
the informations stored in this field, the network informs the Mobile Stations that they
must send a control block in the uplink direction (see the chapter "9.8.4 Relative
Reserved Block Period Field (RRBP)").

The GPRS and EGPRS Mobile Stations can be multiplexed dynamically on the same
PDCH by utilizing the USF. The only problem is that if 8PSK modulation is used in the
downlink blocks (because downlink blocks are related to a EDGE TBF), a GPRS mobile
station multiplexed on the same channel is not able to decode the USF value.
So, the network:
– uses the GMSK modulation, i.e., either CS 1 to CS 4 or MCS 1 to MCS 4, in those
blocks that assign the next uplink radio block, or the next four uplink radio blocks, to
a GPRS mobile station;
– may use the 8PSK modulation for the other blocks.
The dynamic allocation using USF granularity requires that a GPRS Mobile Station is
able to do what the USF in an EGPRS GMSK block. This is enabled by setting stealing

58 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

bits in the EGPRS GMSK blocks to indicate CS4 (see the chapter "4.2.2 EGPRS
Channel Coding"). The coding and interleaving of the USF is done as defined for CS4;
this means that:
– a standard GPRS Mobile Station is able to detect the USF in EGPRS GMSK blocks.
The risk that the rest of the block will be misinterpreted as valid information is
assumed to be low;
– an EGPRS Mobile Station can not differentiate CS4 blocks and EGPRS GMSK
blocks by only looking at the stealing bits. This is however not needed for USF detec-
tion, since the USF is signalled in the same way. Further, assuming that the EGPRS
MS knows if it is in EGPRS or standard GPRS mode, it will only have to try to decode
the remainder of the GMSK blocks in one way in order to determine if they were
aimed for it.
Due to synchronization aspects related to the Mobile Station, if standard GPRS Mobile
Stations are multiplexed on the PDCH, at least one Radio Block every 360 ms on the
downlink direction must use GMSK (i.e., standard GPRS or MCS-1 to MCS-4); this
because every Mobile Station shall receive a radio block at least every 360 ms.

4.3.3 Multiplexing MSs on the same PDCH: Configuration


As it has been described, more than one Mobile Station can be multiplexed, either in the
downlink or in the uplink direction, on the same physical data channel (PDCH):
– in uplink direction up to 7 Mobile Stations can be multiplexed on the same physical
data channel; in fact, up to seven USF values can be used to multiplex Mobile
Stations on the same PDCH in the uplink direction;
– in the downlink direction, up to 16 Mobile Stations can be multiplexed on the same
physical data channel; this number is imposed by the Timing Advance Index (TAI)
necessary for the Timing Advance Update procedure (see the chapter "4.6 Packet
Timing Advance Estimation");
– in total (uplink and downlink) up to 16 Mobile Station can be multiplexed on the same
physical channel; this number is imposed by the Timing Advance Index (TAI)
necessary for the Timing Advance Update procedure (see the chapter "4.6 Packet
Timing Advance Estimation").
If, for instance, 7 Mobile Stations are using a PDCH in the uplink direction, at most
9 Mobile Stations can use the same PDCH but in the downlink direction.
The GMANMSAL attribute allows the user to define the maximum number of users that
can share the same timeslot (PDCH) in uplink (UL) and downlink (DL) direction. It is
composed of two fields:
• the first field specifies the maximum number of users in the uplink direction;
• the second field specifies the maximum number of users in the downlink direction.

4.4 GPRS/EGPRS Logical Channels


Regarding packet switched services (PS) the following Packet Data Logical Channels
have been defined:
• Packet Broadcast Control Channel (PBCCH);
• Packet Common Control Channel (PCCCH);
• Packet Data Traffic Channel (PDTCH);
• Packet dedicated control channels (PTCCH and PACCH).

A30808-X3247-L24-5-7618 59
GPRS/EGPRS Global Description Information
System

These different packet data logical channels, which are specific for the GPRS/EGPRS
technology, can share the same physical channel (on the same PDCH), when the
timeslot is assigned to the GPRS/EGPRS users.
The sharing of the physical channel is based on blocks of 4 consecutive Normal Bursts,
i with the exception of the PTCCH (uplink direction) and the PRACH (see the chapter
:"4.4.4 Packet Dedicated Control Channels") where single Access Bursts are used.

4.4.1 Packet Broadcast Control Channel (PBCCH)


Within the GSM network, system information messages are regularly broadcasted by
the BCCH and busy TCHs. With the help of system information the Mobile Station is able
to decide whether and how it may gain access to the network via the current cell.
The PBCCH logical channel broadcasts Packet data specific System Information (PSI).
In addition to this kind of information, the PBCCH reproduces the information trans-
mitted on the BCCH, to allow circuit switched operation to Mobile Stations that support
both the services. In this way, a MS in GPRS attached mode monitors the PBCCH only,
when this last is configured.
The presence of the PBCCH is not mandatory in a cell supporting packet data (PS)
services; if the PBCCH is not allocated, the packet data specific system information is
broadcast on the BCCH (in the system information 13 message).
The following Packet System Information exists: PSI1, PSI2, PSI 3, PSI3 bis, PSI3 ter,
PSI3 quater, PSI5, PSI13 (see the specification: GSM 04.60):
• PSI1 message is sent by the network on either the PBCCH or PACCH channel,
providing information for cell selection, for control of the PRACH, for description of
control channels and optional power control parameters;
• PSI2 message is sent by the network on the PBCCH channel, providing information
of reference frequency lists, mobile allocations and PCCCH channel descriptions
applicable for the packet access in the cell;
• PSI3 message is sent by the network on the PBCCH or PACCH providing informa-
tion of the BCCH allocation (BA(GPRS)) in the neighbour cells and cell selection
parameters for serving cell and non-serving cells;
• PSI3 bis message is sent by the network on the PBCCH and PACCH providing infor-
mation of the BCCH allocation in the neighbour cells and cell selection parameters
for non-serving cells;
• PSI3 ter message is sent by the network on the PBCCH or PACCH providing infor-
mation on additional measurement and reporting parameters;
• PSI3 quater message is sent by the network on the PBCCH or PACCH providing
information on 3G Neighbour Cells and additional measurement and reporting
parameters;
• PSI5 message contains specific parameter for measurement reporting and network
controlled cell reselection;
• PSI13 message is sent by the network on the PACCH channel (when the PBCCH
channel is not configured), providing the Mobile Station with GPRS/EGPRS cell
specific access-related information (e.g., Page Mode, Routing Area Code, Network
Control Order, Power Control Parameters).
When a GPRS mobile station camps on a cell, it starts reading the system information
on the BCCH channel. From the BCCH channel, the MS understands if the cell supports
the GPRS service or not. If the cell supports the service, the Mobile Station starts
reading the system information 13, that provides GPRS cell specific information. From

60 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

system information 13, the Mobile Station also learns if the PBCCH channel is config-
ured in the cell. If it is configured, the Mobile Station stops reading system information
on the BCCH and starts reading packet system information on the PBCCH.
When an EGPRS mobile station camps on a cell it starts reading system information on
the BCCH channel. From the BCCH channel, it understands if the cell supports the
GPRS service. If the cell supports the service, the Mobile Station starts reading the
system information 13 message that provides information regarding the EGPRS avail-
ability too. From the system information 13 the Mobile Station also learns if the PBCCH
channel is configured in the cell, then:
a) if the PBCCH is not supported, the Mobile Station knows that EGPRS is available
reading GPRS Cell Option IE in the System Information 13 message and finding the
EGPRS Packet Channel Request support indication field. This field indicates if the
EGPRS PACKET CHANNEL REQUEST message is supported in the cell (see for
more details the chapter: "9.8.2.4 TBF Establishment for EDGE Mobile Stations").
Additionally the PSI13 message within the PACCH contains GPRS Cell Options
updated for EGPRS.
b) if the PBCCH is supported, GPRS Cell Options, updated for EGPRS, will be present
in the PSI1 message,
The PBCCH channel, when configured, is allocated on a PDCH physical channel
i (see Fig. 4.10). Only one PDCH can support the PBCCH channel, i.e., it is not possible
to configure the packet system information in two different PDCHs (it is like the GSM
system, where the BCCH channel always resides in the slot 0 of the BCCH TRx).

TDMA frame

BCCH PBCCH

0 7
PDCH

Fig. 4.10 Example of Mapping of the PBCCH Channel.

When, for GPRS/EGPRS mobile stations, the PBCCH is used instead of the BCCH,
i more information and parameters regarding packet switched (PS) data services are
transmitted; for example new cell re-selection criteria are implemented (see the
chapter:"10.1 Cell Selection and Re-selection").

4.4.2 Packet Common Control Channel (PCCCH)


The PCCCH channel comprises logical channels used for common control signalling,
introduced to support packet data services. These common channels are the following:
(see also the "Fig. 4.11 Packet Common Control Channels").
– Packet Paging Channel (PPCH): this channel is used, in the downlink direction
only, to page a Mobile Station prior to the downlink packet transfer. PPCH uses
paging groups in order to allow the usage of DRX mode.
– Packet Access Grant Channel (PAGCH): this channel is used, in the downlink
direction only, in the packet transfer establishment phase, to send resource assign-
ments to Mobile Stations prior to the packet transfer;

A30808-X3247-L24-5-7618 61
GPRS/EGPRS Global Description Information
System

– Packet Random Access Channel (PRACH): this channel is used, in the uplink
direction only, by a Mobile Station to initiate the uplink transfer for sending data or
signalling information. Access Bursts are used on the PRACH channel (see for more
informations the chapter: "4.2 Channel Coding").

Packet Random
- to initiate uplink transfer
Access Channel
- to request allocation of new PDTCHs
PRACH

Packet Common Packet Paging


Control Channels - to page a MS prior of a downlink
Channel
PCCCH transfer
PPCH

Packet Access
Grant Channel - to allocate resources
PAGCH

Fig. 4.11 Packet Common Control Channels

PCCCH channels do not have to be allocated permanently on the cell. Whenever the
PCCCH channels are not allocated, the already configured CCCH channels (i.e.,the
PCH, AGCH and RACH) are used to execute the described operations, in the same way
and with the same GSM functionalities. The existence and the location of the PCCCH
(i.e., the existence and the location of the PDCH channel that support the PCCCH) are
broadcast on the cell.
The first PCCCH channel is automatically allocated when the PBCCH channel is config-
i ured, and it resides in the same PDCH containing also the PBCCH as represented in
the "Fig. 4.12 Example of Mapping of the PCCCH Channel.".
If the user needs more packet common signalling resources, it can configure another
PCCCH in another PDCH as represented in the "Fig. 4.13 Example of Mapping of two
PCCCH Channels.".

TDMA frame
PBCCH
BCCH +
PCCCH
0 7
PDCH

Fig. 4.12 Example of Mapping of the PCCCH Channel.

62 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

TDMA frame
PBCCH
BCCH + PCCCH
PCCCH
0 7
PDCH

Fig. 4.13 Example of Mapping of two PCCCH Channels.

4.4.3 Packet Data Traffic Channel (PDTCH)


The PDTCH is a channel allocated for data transfer. It is temporarily dedicated to one
Mobile Station. In the multislot operation (see the chapter: "4.7 Multislot Configuration"),
one Mobile Station uses multiple PDTCHs in parallel for individual packet transfer (i.e.,
it uses a PDTCH on each assigned PDCH).
All packet data traffic channels are uni-directional:
– uplink PDTCH (PDTCH/U), for a mobile originated packet transfer;
– downlink PDTCH (PDTCH/D) for a mobile terminated packet transfer.
A PDTCH includes also its dedicated control channels (see the chapter "4.4.4 Packet
Dedicated Control Channels").
Regarding the PDTCH assignment, according to the GMANMSAL attribute of the
PTPPKF Managed Object, the following restrictions must be satisfied (see also the
chapter: "4.3.3 Multiplexing MSs on the same PDCH: Configuration").

PDTCHUp<=7
PDTCHDown<=16
PDTCHUp + PDTCHDown <=16

4.4.4 Packet Dedicated Control Channels


Two types of packet dedicated control channels are supported by the GPRS/EGPRS
services:
• Packet Associated Control Channel (PACCH):The PACCH channel conveys
signalling information related to a given Mobile Station. The signalling information
includes for example acknowledgements and power control information. PACCH
also carries resource assignment and reassignment messages, comprising the
assignment of resources for PDTCH(s) and for further occurrences of PACCH.
The PACCH channel is always associated to PDTCH channels: the PDTCH channel
allows transmission of data blocks, while the PACCH channel allows transmission of
the related signalling blocks; so the PACCH channel shares with PDTCHs the
resources that have been currently assigned to one Mobile Station.
For instance, the following control messages can be carried by the PACCH channel,
according to the direction of transmission:
Uplink Direction (UL):
– Packet Control Acknowledgment;
– Packet Downlink Ack/Nack.

A30808-X3247-L24-5-7618 63
GPRS/EGPRS Global Description Information
System

Downlink Direction (DL):


– Packet Uplink Ack/Nack;
– Packet Power Control/timing Advance.
• Packet Timing Advance Control Channel (PTCCH): this type of channel is used
in the timing advance update procedure (see "4.6 Packet Timing Advance Estima-
tion"). The PTCCH/U is used to transmit the random access burst to allow the esti-
mation of the timing advance for one Mobile Station in packet transfer mode; the
PTCCH/D is used to transmit timing advance information updates to several Mobile
Stations.

4.4.5 Coding of GPRS/EGPRS Logical Channels


Regarding the coding of packet switched logical channels, the following considerations
are necessary:
• Packet Data Traffic channels (PDTCHs) use:
– coding schemes from CS1 to CS4 in cases of support of the GPRS technology;
– coding schemes from MCS1 to MCS4 in cases of support of the EGPRS tech-
nology with GMSK modulation;
– coding schemes from MCS5 to MCS9 in cases of support of the EGPRS tech-
nology with 8PSK modulation;
• for all the packet control channels, with the exception of the PRACH and PTCCH/U,
the CS1 coding scheme is always used;
• both PRACH and PTCCH/U use access bursts; for access bursts, two coding
schemes are specified (8 bit coding and 11 bit coding, see the chapter: "9.8.2.1 8
Bit or 11 Bit Uplink Access").
• PTCCH/D is coded using the CS1 coding scheme.

4.5 Mapping of Logical Channels onto Physical Channels


A physical channel allocated to carry packet logical channels is called a Packet Data
channel (PDCH). A PDCH channel carries packet logical channels only.
The GPRS/EGPRS logical channels are mapped dynamically onto a 52-multiframe See
the "Fig. 4.2 Multiframe Structure for a PDCH".
The 52-multiframe consists of 12 blocks of 4 consecutive frames, 2 idle frames and 2
frames used for the PTCCH channel.
A block allocated to a given logical channel comprises one radio block or, in the uplink
direction only, 4 random access bursts.
Sometimes to allow the estimation of the timing advance, instead of transmitting a radio
i block, four access bursts are sent (see the chapter: "9.8.5 Polling Procedures").
It must be noted that the PRACH channel, when it is used in the uplink direction, it’s not
composed of a single block of four frames, but it is composed of four access bursts.

The type of channel may vary on a block by block basis. From the configuration point of
view the 12 blocks are put in an ordered list, defined as: “B0, B6, B3, B9, B1, B7, B4,
B10, B2, B8, B5, B11.”
A single PDCH carries different logical channels, according to either the configuration’s
actions done by the user or to the direction of transmission. The following configuration
can be used for a single PDCH:

64 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

a) the PDCH does not carry the specific GPRS/EGPRS signalling (for example PBCCH
and PCCCH channels);
b) the PDCH carries both PBCCH and PCCCH channels;
c) the PDCH carries GPRS/EGPRS common signalling (for example PCCCH) but not
the PBCCH channel.

4.5.1 PDCH without the Specific GPRS/EGPRS Signalling


When the PDCH does not carry GPRS/EGPRS specific signalling, the following state-
ments are significant:
• in the downlink direction (see the "Fig. 4.14 Example of Mapping of Logical Chan-
nels in the Physical Channel (Downlink Direction).") all blocks can be used as
PDTCH/D or PACCH/D: the logical channel type is indicated in the block header. The
mobile owner of the PDTCH/D or PACCH/D is indicated by the parameter TFI
(Temporary Flow Identifier);

PDCH Multiframe (downlink direction)

PDTCH/D PACCH/D PACCH/D PDTCH/D PDTCH/D PDTCH/D

Fig. 4.14 Example of Mapping of Logical Channels in the Physical Channel


(Downlink Direction).

• in the uplink direction (see the "Fig. 4.15 Example of Mapping of Logical Channels
in the Physical Channel (Uplink Direction).") all blocks can be used as PDTCH/U or
PACCH/U: the occurrence of the PDTCH/U (and/or the PACCH/U) at the given
block(s) Bx (where Bx = B0...B11) in the 52-multiframe structure for a given Mobile
Station on a given PDCH is indicated by the value of the Uplink State Flag (USF).
The USF is contained in the header of the preceding block, transmitted in the down-
link of the same PDCH.
The Mobile Station may transmit a PDTCH block or a PACCH block on any of the
uplink blocks used for the purpose.
The occurrence of the PACCH/U associated to a PDTCH/D is indicated by the
network by polling the Mobile Station to transmit the PACCH/U block (as described
in the chapter:"9.8.4 Relative Reserved Block Period Field (RRBP)".

PDCH Multiframe (uplink direction)

PDTCH/U PDTCH/U PACCH/U PDTCH/U PACCH/U PDTCH/U

Fig. 4.15 Example of Mapping of Logical Channels in the Physical Channel


(Uplink Direction).

4.5.2 PDCH Carrying both PBCCH and PCCCH


When the PDCH carries both the PBCCH and the PCCCH channels, the following state-
ments are significant:

A30808-X3247-L24-5-7618 65
GPRS/EGPRS Global Description Information
System

a) DOWNLINK DIRECTION:
– the first block (B0) of the multiframe (see the "Fig. 4.2 Multiframe Structure for a
PDCH") is reserved for the PBCCH channel; the user can also configure up to 3
more blocks as additional PBCCH. To configure additional blocks as PBCCH
blocks, the BSPBBLK attribute can be configured by the user: this attribute allows
the specification of at most four blocks, following the order: B0, B6, B3, B9.
– the next remaining blocks can be configured for PAGCH, PDTCH/D and
PACCH/D. To configure additional blocks to carry PAGCH, PDTCH/D and
PACCH/D, the BPAGCHR attribute can be configured by the user. This attribute
allows the specification of at most 12 blocks, following the order: B6, B3, B9, B1,
B7, B4, B10, B2, B8, B5, B11.
– the remaining blocks are used for PPCH, PAGCH, PDTCH/D and PACCH/D.
The "Fig. 4.16 Example of Downlink Configuration with PBCCH and PCCCH Chan-
nels" shows an example of one PDCH carrying both the PBCCH and the PCCCH
channels, where 3 blocks are dedicated to the PBCCH channel by setting the value
of the BSPBBLK attribute to 2. It can be noted how the number of blocks assigned
to the logical channels change according to the value of the BSPBBLK attribute.
In this example, since three blocks are always dedicated to the PBCCH channel, at
most 9 blocks can be dedicated to the PAGCH channel by the BPAGCHR parameter.

b) UPLINK DIRECTION:
– in the uplink direction, each block can be used as PRACH, PDTCH/U and
PACCH/U; the BPRACHR attribute allows the user to indicate how many blocks
must be reserved in a fixed way to the PRACH channel. The user can reserve up
to 12 blocks (i.e., all the multiframe) for the PRACH channel. Remember that in a
PRACH block, 4 random access bursts are always sent.
The "Fig. 4.17 Example of Uplink Configuration with PRACH Channel." shows an
example of one PDCH carrying PRACH channel; note how the blocks assigned to
the logical channels change according to the value given to the BPRACHR attribute.

66 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

BSPBBLK BPAGCHR BO B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 B11

2 0
2 1
2 2
2 3
2 4
2 5
2 6
2 7
2 8
2 9
2 :

PBCCH

PAGCH + PDTCH + PACCH

PAGCH + PDTCH + PACCH + PPCH

Fig. 4.16 Example of Downlink Configuration with PBCCH and PCCCH Channels

BPRACHR BO B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 B11

0
1
2
3
4
5
6
7
8
9
10
11
12

PRACH

PRACH + PDTCH + PACCH

A30808-X3247-L24-5-7618 67
GPRS/EGPRS Global Description Information
System

Fig. 4.17 Example of Uplink Configuration with PRACH Channel.

4.5.3 PDCH Carrying PCCCH


When the PDCH carries the PCCCH channel (without the PBCCH one), the following
statements have to be considered:
a) DOWNLINK DIRECTION:
– up to 12 blocks can be configured for PAGCH, PDTCH/D and PACCH/D; to
configure blocks to carry PAGCH, PDTCH/D and PACCH/D, the BPAGCHR
attribute can be configured by the user. This attribute allows the specification of
at most 12 blocks, following the order: B0, B6, B3, B9, B1, B7, B4, B10, B2, B8,
B5, B11.
– the remaining blocks are used for PPCH, PAGCH, PDTCH/D and PACCH/D.
b) UPLINK DIRECTION:
– in the uplink direction each block can be used as PRACH, PDTCH/U and
PACCH/U; the BPRACHR attribute allows the user to indicate how many blocks
must be reserved in a fixed way to the PRACH channel. The user can reserve up
to 12 blocks (i.e., all the multiframe) for the PRACH channel. It is important to
outline that in a PRACH block, 4 random access bursts are always sent.

4.6 Packet Timing Advance Estimation


The packet timing advance estimation procedure is used to derive the correct value for
timing advance that the Mobile Station has to use for the uplink transmission of radio
blocks.
The timing advance procedure is organized into two parts:
– initial timing advance estimation;
– continuous timing advance update.

4.6.1 Initial Timing Advance Estimation


The initial timing advance estimation is based on the single access burst carrying the
Packet Channel Request. The Packet Uplink Assignment or Packet Downlink Assign-
ment (see the chapter: "9.8 Access to the Network (Establishment of a TBF)") then
carries the estimated timing advance value to the Mobile Station. This value is used for
the uplink transmissions until the continuous timing advance update provides a new
value (see the chapter: "4.6.2 Continuous Timing Advance Update").
When Packet Downlink Assignment has to be sent without prior paging (i.e., when the
MS in the Ready state, see "9.3.1 Mobility Management States"), no valid timing
advance value may be available.
When timing advance information is not provided in the assignment message, the
mobile station is not allowed to send normal bursts on the uplink direction until it receives
a valid timing advance value.
To get a valid timing advance value, the continuous timing advance update procedure
has been introduced by the specification (see the chapter: "4.6.2 Continuous Timing
Advance Update"). Since this procedure could create some delays between the packet
downlink assignment message and the beginning of the data transfer in downlink direc-
tion, in order to reduce the time between a packet downlink assignment message and

68 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

the effective beginning of downlink data transmission, a specific polling procedure is


executed to get the timing advance value.
This polling procedure is basically the following (The chapter: "9.8.5 Polling Procedures"
describes more details about it):
1. With the Packet Downlink Assignment message the network polls the MS to send a
Packet Control Acknowledgment message formatted as four access bursts.
2. The network calculates the initial timing advance value from these access bursts.
3. The network, by means of the PACKET TIMING ADVANCE/POWER CONTROL
message, sends to the MS the calculated timing advance value.

4.6.2 Continuous Timing Advance Update


The Mobile Stations in Packet transfer mode use the continuous timing advance update
procedure.
This procedure is carried on the PTCCH allocated to the Mobile Station. The following
statements have to be considered:
• For the uplink packet transfer within the Packet Uplink Assignment the network
assigns to the Mobile Station (besides the USF and the TFI) the parameter: Timing
Advance Index (TAI) and a PTCCH channel.
• For the downlink packet transfer, within the Packet Downlink Assignment, the
network assigns to the Mobile Station (besides the TFI) the parameter: Timing
Advance Index (TAI) and a PTCCH channel.
The TAI parameter specifies the PTCCH channel that the Mobile Station will use.
On the uplink, the Mobile Station sends an access burst in the assigned PTCCH/U,
which is used by the network to derive the timing advance value.
The network analyzes the received access bursts and determines a new timing advance
value for all the Mobile Stations performing the continuous timing advance update
procedure on that PDCH. The new timing advance value is sent via a downlink signalling
message (TA-message) on PTCCH/D. The network can also send the timing advance
information within the Packet Timing Advance/Power Control and Packet Uplink
Ack/Nack messages on the PACCH.
The "Fig. 4.18 Continuous Timing Advance Update Feature" shows the mapping of
both the uplink access bursts and the downlink TA-messages on groups of eight 52-
multiframes:
• the TAI value shows the position where a slot is reserved for a Mobile Station to send
an access burst in the uplink direction (e.g., TAI= 1 identifies the multiframe number
n and the idle frame number 2). The TAI value defines the used PTCCH/U sub-
channel, e.g.,
– the MS with TAI= 1 sends its access burst every eight multiframes in the idle
frame number 2;
– the MS with TAI= 5 sends its access burst every eight multiframes in the idle
frame number 10;
• the TA-message is sent from the network to Mobile Stations and it is composed of
a radio block sent over four frames.
For example:
– the Mobile Stations that have sent their access bursts in idle frames number 0, 2,
4 and 6, will receive their TA-message (with information about the timing advance
to be used) in the TA-message number 2. This TA-message is transmitted in the

A30808-X3247-L24-5-7618 69
GPRS/EGPRS Global Description Information
System

downlink direction in terms of a radio block, distributed on the frames number 8,


10, 12 and 14.
– the Mobile Stations that have sent their access bursts in idle frames number 8,
10, 12 and 14, will receive their TA-message (with information about the timing
advance to be used) in the TA-message number 3. This TA-message is trans-
mitted in the downlink direction in terms of a radio block, distributed on the frames
number 16, 18, 20 and 22.

Fig. 4.18 Continuous Timing Advance Update Feature

The continuous timing advance update procedure could create some delays between
i the packet downlink assignment message and the beginning of the data transfer in the
downlink direction. In order to reduce the time between a packet downlink assignment
message and the effective beginning of downlink data transmission, a specific packet
polling procedure has to be executed (see the chapter: "9.8.5 Polling Procedures" for
more details about the procedure).

4.7 Multislot Configuration


Multiple packet data traffic channels can be allocated to the same Mobile Station. This
is referred to as the “multislot packet” configuration. In this configuiration one Mobile
Station may use multiple PDTCHs in parallel for individual packet transfer.
The network may allocate to a Mobile Station:
– several PDTCH/Us for one mobile originated communication.
– several PDTCH/Ds for one mobile terminated communication.
In this context, allocation refers to the list of PDCHs that may dynamically carry the
PDTCHs for that specific Mobile Station. The PACCH may be mapped onto any of the
allocated PDCHs.

70 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

When a multislot configuration is used, a certain number of timeslots (PDCHs) are allo-
cated to the same Mobile Station, accordingly to its multislot capability; the following
rules must be satisfied when more than one timeslot is assigned:
1. timeslots must belong to the same frequency (i.e., to the same TRX).
2. timeslots must be adjacent (i.e., they must have neighboring timeslots numbers-TN).
3. timeslots must belong to the same frequency hopping law, i.e., they must have the
same:
– Mobile Allocation (MA);
– Mobile Allocation Index Offset (MAIO);
– Hopping Sequence Number (HSN);
Regarding frequency hopping for GPRS/EGPRS services, both Base Band
i Frequency Hopping and Synthesizer Frequency Hopping are supported.

4. timeslots shall have the same training sequence code (TSC).

4.7.1 Mobile Station Classes for Multislot Capabilities


When a GPRS/EGPRS Mobile Station supports the configuration of multiple timeslots,
it will belong to a multislot class as defined in the Tab. 4.6 (only the first 12 classes are
represented in the table):

Multislot Maximum number of slots Minimum number of slots


class
Rx_max Tx_max Sum Ttb Tra

1 1 1 2 2 4
2 2 1 3 2 3
3 2 2 3 2 3
4 3 1 4 1 3
5 2 2 4 1 3
6 3 2 4 1 3
7 3 3 4 1 3
8 4 1 5 1 2
9 3 2 5 1 2
10 4 2 5 1 2
11 4 3 5 1 2
12 4 4 5 1 2

Tab. 4.6 Mobile Station Multislot Classes

where the fields: “Rx_max, Tx_max, Sum, Ttbm, Tra” have the following meaning:
– Rx_max describes the maximum number of timeslots that the Mobile Station can
use per TDMA frame in the downlink direction. It shall be able to support all integer
values of timeslots (from 0 to Rx_max) in the downlink direction;
– Tx_max describes the maximum number of timeslots that the Mobile Station can
use per TDMA frame in the uplink direction. It shall be able to support all the integer
values of timeslots (from 0 to Tx_max) in the uplink direction;

A30808-X3247-L24-5-7618 71
GPRS/EGPRS Global Description Information
System

– Sum is the total number of uplink(Tx) and downlink(Rx) timeslots that the Mobile
Station can use per TDMA frame (when it has established a TBF in both the direc-
tions). The MS must be able to support all combinations of integer values of Rx and
Tx timeslots where; 1 <= Rx + Tx <= Sum , Rx<=Rx_max and Tx<=Tx_max;
– Ttb relates to the time needed for the Mobile Station to get ready to transmit. This
minimum requirement is used when adjacent cell power measurements are not
required by the service selected. For type 1 Mobile Station it is the minimum number
of timeslots that will be allowed between the end of the last previous received
timeslot and the first next transmit timeslot or between the previous transmit timeslot
and the next transmit timeslot when the frequency is changed in between;
– Tra relates to the time needed for the Mobile Station to perform adjacent cell signal
level measurement and get ready to receive after it has transmitted in the uplink
direction. For type 1 MS it is the minimum number of timeslots that are allowed
between the previous transmit or receive timeslot and the next receive timeslot when
measurement is to be performed between. For type 2 MS it is the minimum number
of timeslots that are allowed between the end of the last receive burst in a TDMA
frame and the first receive burst in the next TDMA frame.
Type 1 Mobile Station are not required to transmit and receive at the same time,
i whereas Type 2 Mobile Station are instead required.

The "Fig. 4.19 Example of Multislot Configuration" shows an example regarding a


Mobile Station with multislot class=8: the MS has established concurrent TBFs, and it
has 4 timeslots in the downlink direction and 1 timeslot in the uplink direction (between
downlink and uplink TDMA frames there is a temporal offset of 3 timeslots). When the
Mobile Station has monitored its last downlink timeslot, it shall wait Ttb timeslots (i.e., 1
timeslot) before transmitting; after having transmitted in the uplink direction, it shall wait
Tra timeslots (i.e., 2 timeslots) before starting to monitor on the downlink direction.

Mobile Class = 8
Rx = 4
Tx = 1
Sum = 5
Ttb= 1
Tra = 2
TDMA frame - Downlink
0 7 0
d d d d d d

u 7

0 7
TDMA frame - Uplink

Ttb Tra

Fig. 4.19 Example of Multislot Configuration

4.7.2 Mapping of Uplink Packet Traffic Logical Channels


The PDCHs where the Mobile Station may expect occurrence of its PDTCH/U(s) or
PACCH/U for a originated transfer is indicated in the resource allocation messages. The

72 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

PACCH/U is allocated respecting the resources allocated to the Mobile Station and to
its multislot class.
For each PDCH allocated to the Mobile Station, an USF value is assigned to it.
To establish a multislot uplink Temporary Block Flow, the following conditions shall be
satisfied:
– one common uplink TFI is available in all timeslots (TFI is the same on all the
PDCHs).
– one USF is available for each PDCH in the set.
– a TAI is available in one of the assigned timeslots.

4.7.3 Mapping of Downlink Packet Traffic Logical Channels


The PDCHs where the Mobile Station may expect occurrence of its PDTCH/D(s) or
PACCH/D for a mobile terminated transfer is indicated within the resource allocation
messages.
The logical channel type is indicated in the block header. The Mobile Staiton owner of
the PDTCH/D or PACCH/D is indicated by the related TFI parameter (Temporary Frame
Identifier).
To establish a multislot downlink Temporary Block Flow, the following conditions shall
be satisfied:
– one common downlink TFI is available in all timeslots (TFI is the same on all the
PDCHs).
– a TAI is available in one of the assigned timeslots.

A30808-X3247-L24-5-7618 73
GPRS/EGPRS Global Description Information
System

5 Radio Resources Management


The PTPPKF functional object represents point to point packet data services in a cell.
Point to point data services regard:
• GPRS service, i.e., packet switched data services using the traditional GMSK modu-
lation;
• EGPRS service: EGPRS introduces the 8PSK modulation in the existing GSM
network.
The creation of the PTPPKF object for a specific cell can be done only if:
– the BTS object has been already created for the cell;
– at least one PCU (see, "6 Hardware and Software Architecture") has been created;
– at least one permanent virtual connection (see "7 Gb Interface") has been config-
ured for the PCU.
Either during or after having enabled GPRS/EGPRS services in a cell, the operator can
configure, according to his needs, the radio resources of the cell.
A cell supporting GPRS/EGPRS may allocate resources on one or several physical
channels in order to support the packet switched data traffic. The physical channels (i.e.,
PDCHs), shared by GPRS/EGPRS MSs, are taken from the common pool of physical
channels available in the cell. The allocation of physical channels to circuit switched
services and packet switched services is done according to the following two principles:
• Master-slave concept: at least one PDCH, acting as a master, accommodates:
– the packet broadcast channel (PBCCH);
– the packet common control channel (i.e., PCCCH);
– user data and dedicated signalling (i.e., PDTCH and PACCH).
Other PDCHs, acting as slaves, are used for both user data transfer and dedicated
signalling only (PACCH).
The master-slave concept is only valid when control signalling is carried on PCCCH;
if controllo signalling is carried by the existing CCCH, the concept is no longer valid.
• Capacity on demand concept: GPRS/EGPRS does not require permanently allo-
cated PDCHs. The allocation of capacity for these services can be based on the
needs for actual packet transfers which is here referred to as the "capacity on
demand" principle.
The operator can, as well, decide to dedicate permanently or temporarily some
physical resources (i.e., PDCHs) to the packet switched data traffic
When the PDCHs are congested due to the GPRS/EGPRS traffic load and more
resources are available in the cell, the network can allocate more physical channels
as PDCHs. However, the existence of PDCH(s) does not imply the existence of
PBCCH/PCCCH.
When no PBCCH/PCCCH channels are allocated in a cell, all GPRS/EGPRS attached
i MSs camp on the BCCH/CCCH. When PCCCH is allocated in a cell, all GPRS/EGPRS
attached MSs camp on it.
PCCCH can be allocated either as the result of the increased demand for packet data
transfers or whenever there are enough available physical channels in a cell (to
increase the quality of service). When the PCCCH capacity is inadequate, it is possible
to allocate additional PCCCH resources on one or several PDCHs (see "6.4 Packet
Switched Services Supported on CCCH/PCCCH").

74 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

According to previous concepts, the operator can use different strategies to configure
packet switched data services in a cell; e.g., he can:
a) reserve at least one static timeslot for GPRS/EGPRS specific signalling, and
configure other dynamic timeslots (which will be shared with circuit switched
services) for GPRS/EGPRS data;
b) reserve at least one timeslot for GPRS/EGPRS specific signalling, and configure
other static timeslots (which will not be shared with circuit switched services) for
GPRS/EGPRS data;
c) not reserve any timeslot for GPRS/EGPRS specific signalling, and configure some
static timeslots for GPRS/EGPRS data;
d) not reserve any timeslot for GPRS/EGPRS specific signalling, and configure some
dynamic timeslots for GPRS/EGPRS data;
e) not reserve any timeslot for GPRS/EGPRS specific signalling, and configure both
some static and some dynamic timeslots.

5.1 Enabling Packet Switched Services in a Cell


When configuring GPRS/EGPRS services on a cell (i.e., when creating a PTPPKF
object instance), the user can specify some parameters (belonging to the PTPPKF
object) that allow him to configure the two services according to his needs.
Among the parameters of the PTPPKF object, two of them allow making available or not
packet data services in the referred cell. In fact, when a PTPPKF object instance is
created, neither GPRS service nor EGPRS service are enabled as default. So the
user, to enable GPRS and EGPRS, has to manage two specific parameters; they are:
1. the EGPRS flag (enable GPRS) allows to make available or not available the whole
Packet Data services in the cell, i.e., both GPRS (based on CS1,.., CS4), and
EGPRS (based on MCS1,.., MCS9). The EGPRS flag has the same functional
behavior as the command lock/unlock on the PTPPKF object; the only difference is
that, in cases of EGPRS=FALSE, the operational state of PTPPKF is disabled. In
this way, the operator is clear that the PTPPKF cannot provide service, even if it
unlocked.
The default value of the EGPRS flag is FALSE; this means that when creating the
PTPPKF object instance, the cell is disabled to support, in general, both (GPRS and
EGPRS) packet data services. But it must be noted that:
– before setting EGPRS=TRUE for a specific cell, the user will specify, for this cell,
at least one TRX that supports the GPRS service; i.e., he can choose, among the
total number of TRXs configured for the cell, which of them will handle packet data
services (see "5.1.1 Enabling GPRS Service in the Cell").
– to enable EGPRS services in the cell is not sufficient to set EGPRS=TRUE but
other parameter settings are required (see below); but setting EGPRS=TRUE is
the first step in the procedure that allows the configuration of the EGPRS service
in the cell (see "5.1.2 Enabling EGPRS Service in the Cell").
2. the EEDGE (enable EDGE) flag allows making available or not available the
EGPRS service in the cell, provided that the GPRS service is available and there
are radio resources configured to support EDGE. It is not allowed to make the
EGPRS service available (EEDGE=TRUE) if the GPRS service is not available
(EGPRS=FALSE) or if no TRX in the cell has been configured to support EDGE.
Moreover, the EEDGE flag cannot be set to TRUE if CSCH3CSCH4SUP parameter
is set to FALSE, meaning that the EGPRS service can be actiivated if the CS-3/CS-
4 coding schemes are enabled.

A30808-X3247-L24-5-7618 75
GPRS/EGPRS Global Description Information
System

Nevertheless, the EGPRS service can be also activate without activation of CS-
3/CS-4 (see note below).
To make available the EGPRS service without activation of CS-3 /CS-4 coding
i schemes, the operator shall set the bit25 of MNTBMASK parameter to TRUE,meaning
that the max coding scheme usable will be CS-2 independently from
CSCH3CSCH4SUP value set to TRUE.Other parameter settings are required to enable
EGPRS services in the cell(see "5.1.2 Enabling EGPRS Service in the Cell").The
default value of the EEDGE flag is FALSE.

5.1.1 Enabling GPRS Service in the Cell


As it has been described in "5.1 Enabling Packet Switched Services in a Cell", when a
PTPPKF object instance is created, the EGPRS parameter, that allows making available
or not available the whole Packet Data services in a cell (both GPRS and EGPRS), is
set to FALSE as a default value.
Before setting the EGPRS parameter to true, the user must enable the GPRS service
on a TRX basis, i.e., the user must enable at least one TRX to support packet data
services.
The GPRS service will not be enabled in a cell, if the user does not enable it on at least
! one TRX of the cell, after having created the PTPPKF object.

To enable, in general packet switched data services on a specific TRX, the user must
set to “TRUE” the GSUP attribute belonging to the TRX object.
Setting GSUP=TRUE means that the TRX is enabled to support, in general, both packet
data services, i.e., it is enabled to support both GPRS and EGPRS.
Fig. 5.1 shows an example of one cell with five TRXs, where three of them (i.e., TRX0,
TRX3 and TRX4) have been enabled to support the packet switched services.

76 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

TDMA frame

TRX 0 BCCH GSUP=TRUE

0 7
TDMA frame

TRX 1 GSUP=FALSE

0 7
TDMA frame

TRX 2 GSUP=FALSE

0 7
TDMA frame

TRX 3 GSUP=TRUE

0 7

TDMA frame

TRX 4 GSUP=TRUE

0 7

Fig. 5.1 Example of TRXs enabled to support Packet Switched Services.

BTSplus equipment can be equipped with two different types of carrier units (see 6.2):
1. GSM-CUs, i.e., carrier units able to support GSM and GPRS services;
2. E-CUs (EDGE carrier units), i.e., carrier units able to support GSM, GPRS and
EGPRS services.
The user, beside enabling the TRX to support packet data services (with GSUP param-
eter), can indicate if the TRX will be used for GSM/GPRS only, or if it will be used for
EGPRS too. To indicate how the TRX has to be used, the TRXMD parameter is used; it
can assume two values:
– GSM: it is the default value that means that the TRX can be used for GSM, and also
for GPRS if GSUP=TRUE;
– EDGE: it means that the TRX can be used for GSM, and also for both GPRS and
EGPRS if GSUP=TRUE.
Regarding the detailed procedure to enable EGPRS, please refer to "5.1.2 Enabling
i EGPRS Service in the Cell".

From the BTS equipment point of view, the TRXMD parameter is the criterion used to
i allocate a carrier unit type (GSM-CU or E-CU) to the transceiver. The association
between a TRX and the boards (CU or E-CU) of a BTSplus is performed automatically
by the BTS equipment, taking into account suggestion from the operator (i.e., the
TRXMD setting).

A30808-X3247-L24-5-7618 77
GPRS/EGPRS Global Description Information
System

After having enabled one or more TRXs to support GPRS, i.e., after having set for one
or more TRXs:
– GSUP=TRUE
– TRXMD=GSM
the user must set to TRUE the EGPRS parameter of the PTPPKF object, to definitely
enable the GPRS service in the cell.
It is possible to set GSUP =TRUE for whatever TRX of the cell, with the following excep-
tions:
– In cases of Concentric Cells, the TRXs supporting GPRS must always belong to the
complete cell area;
– In cases of GSMDCS (common BCCH) cells, all the TRXs that support GPRS must
belong to the same band of the BCCH TRX (and this coincides with the band of the
Complete Area). This is due to the fact that the two GSM and DCS bands have
different propagation factors, thus it could be that on cell borders only the frequency
of one band is received; one mobile that accessed the cell with one band could not
work with the other one;
– In cases of cells having SYSID=EXT900, only the TRXs with TRXFREQ belonging
to BB900 band (that is, the same band of BCCH) can have GSUP=TRUE;
– In cases SYSID=F2ONLY, if the BCCH belongs to BB900 band, all the TRXs for
which the GSUP is set at TRUE must belong to the BB900 band;
– In cases SYSID=F2ONLY, if the BCCH belongs to the extended band, all the TRXs
for which the GSUP is set at TRUE must belong to the extended band;
– It is possible to set as first GPRS TRX any TRX of the cell, that is: it is not mandatory
to set this attribute to TRUE first on the BCCH TRX;
– The setting of a TRX to GSUP = TRUE has to take into account the Multislot
constraints for TSC and Frequency Hopping parameters;
– The setting of a TRX to GSUP=TRUE must be executed only when all the TRX’s
channels are not available to the service (this situation can be reached by executing
a shutdown for all these TCHs: this is suggested to avoid impacts on CS calls.
Beside the TRXs of a cell on which the user wants to configure the packet switched data
i services, it is suggested to also configure GSUP =TRUE for the BCCH TRX. In this way
the condition of no TRXs with GSUP =TRUE (condition that puts the PTPPKF object in
DISABLE state) also happens when there is a BCCH outage. In this case, the whole
BTS is put Out of Service from both circuit switched and packet switched services point
of view.

Once one or more TRXs have been enabled to support the GPRS service, the user can
configure, according to his needs, some static and dynamic GPRS channels on them
(see "5.2 Configuration of GPRS Channels in a Cell").

5.1.2 Enabling EGPRS Service in the Cell


As it has been described in "5.1.1 Enabling GPRS Service in the Cell", to enable GPRS
on a cell and TRX basis, the user must set:
– the GSUP parameter of the TRX object to TRUE (for each TRX he wants to enable
support of packet data services);
– the EGPRS parameter of the PTPPKF object to TRUE;
After that if the user wants to enable EGPRS service in one or more cells of the BSC,
he must execute a series of operations to enable it.

78 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

The ESUP parameter of the BSC object allows the user to enable EGPRS services in
the whole BSC. If the user does not set to TRUE the ESUP parameter, EGPRS service
will not be allowed in the BSC. Since EGPRS can only be enabled on high capacity
BSCs, the ESUP parameter can not be set to true if the NTWCARD parameter (BSC
object) is not set to either SNAP or SNAP_STLP.
After the user has enabled EGPRS service on a BSC basis, he has to enable it both on
a TRX basis and on a cell basis.
As described in "5.1.1 Enabling GPRS Service in the Cell", to enable, in general, packet
data services on a specific TRX, the user must set to “true” the GSUP attribute belonging
to the TRX object.
Setting GSUP =TRUE means that the TRX is enabled to support in general both packet
data services, i.e., it is enabled to support both GPRS and EGPRS.
To activate EGPRS service on a specific TRX, beside enabling the TRX to support
packet data services (with the GSUP parameter), the user has to indicate that the TRX
will also be used for EGPRS. To indicate how the TRX has to be used, the TRXMD
parameter is used; it can assume two values:
– GSM: it is the default value that means that the TRX can be used for GSM and also
for GPRS;
– EDGE: it means that the TRX can be used for GSM and also for both GPRS and
EGPRS.
To enable EGPRS service on a TRX the user must set the TRXMD parameter to the
EDGE value.
After having enabled EGPRS on a TRX basis, i.e., after having enabled at least one
TRX to support EGPRS, the user can enable EGPRS on a cell basis by setting to TRUE
the EEDGE parameter.
Fig. 5.2 shows an example of one cell with five TRXs, where two of them (i.e., TRX0,
TRX3) have been enabled to support GPRS only and one of them (TRX5) has been
enable to support a EGPRS too.

A30808-X3247-L24-5-7618 79
GPRS/EGPRS Global Description Information
System

TDMA frame
GSUP=TRUE
TRX 0 BCCH
TRXMD=GSM
0 7
TDMA frame

TRX 1 GSUP=FALSE

0 7
TDMA frame

TRX 2 GSUP=FALSE

0 7
TDMA frame
GSUP=TRUE
TRX 3 TRXMD=GSM
0 7

TDMA frame
GSUP=TRUE
TRX 4 TRXMD=EDGE
0 7

Fig. 5.2 Example of TRXs enabled to support GPRS and EGPRS.

Once one or more TRXs have been enabled to support the EGPRS service, the user
can configure, according to his needs, some static and dynamic channels on them, to
be used for packet switched services (see "5.2 Configuration of GPRS Channels in a
Cell").

5.1.3 Aspects Related to Carrier Configuration


As described before, by means of the GSUP and TRXMD attributes, it is possible to
specify whether a TRX supports:
– GSM services;
– GPRS and/or EGPRS services
From the BTSE point of view, the EDGE capability of a Carrier Unit(CU) is modelled with
a read only attribute called TRX CAPABILTY. It reflects the real capabilities of the TRX
independently of the TRXMD parameter setting. The attribute is available for the oper-
ator by GETINFO TRX and GETINFO BTS commands, and it can assume the following
values:
– GSM: the TRX is associated to a carrier Unit without EDGE capability;
– EDGE: the TRX is associated to a carrier Unit with EDGE capability;
– UNKNOWN: the BSC has no knowledge about the carrier Unit associated to TRX.
The following combinations of TRXMD and TRX CAPABILITY are possible
(see Tab. 5.1):

80 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

TRXMD TRX CAPABILTITY Meaning

GSM GSM The TRX is working with GSM functionality


GSM EDGE The TRX is working with GSM functionality
GSM UNKNOWN No CU is related
EDGE GSM The TRX is working with GSM functionality
since no E-CU is available
EDGE EDGE The TRX is working with EDGE functionality
EDGE UNKNOWN No CU is related

Tab. 5.1 Combinations of TRXMD and TRX CAPABILITY Values

TRX-CU assignment procedure


The BTSE shall try to find an optimal allocation between the required TRX operation
modes and the available CU types according to the following rules:
1. Try to allocate the BCCH TRX to the appropriate CU type;
2. Try to allocate all EDGE-TRX to E-CUs
3. Try to allocate all GSM-TRX to GSM-CUs
4. Try to allocate all remaining GSM-TRX to E-CUs
5. Try to allocate all remaining EDGE-TRX to GSM-CUs (the state of the TRX changes
to enabled-degraded).
Reconfiguration of TRXs due to a change of Functional Configuration
When TRXs are reconfigurated, the BTSE checks if the allocation among TRXs and
CUs is still optimal. If necessary, an automatic reconfiguration according to the rules
shown above is performed.
A change of the functional configuration may be caused by:
– creation/deletion of a TRX;
– modification of the TRXMD attribute of a TRX;
– breakdown of CU allocated to BCCH-TRX;
– breakdown of CU allocated to complete area TRX in case of concentric cells;
– commissioning of TRX after breakdown.
In some cases, the changes of TRX configuration may lead to a loss of traffic for one or
two TRX. It is an operator’s task to avoid the loss of traffic by taking the right measures.
Therefore the appropriate configuration procedures have to be performed: Shut-
down/Create/Unlock command sequence.

If the BCCH-TRX is involved in the TRX-CU assignment procedure(it can happen,if a


! EDGE property of the BCCH-TRX itself is modified or if the BTSE-internal optimization
algorithm touches the BCCH-TRX), the whole Cell involved.In this case, a shut-
down/create/unlock procedure should be applied to the whole cell and not only to the
single TRX involved.

A30808-X3247-L24-5-7618 81
GPRS/EGPRS Global Description Information
System

In case of Baseband frequency hopping all TRXs involved in the same hopping law must
i be homogeneus, i.e. they must have the same TRXMD value. If a TRX with TRXMD =
EDGE gets TRX_CAPABLITY = GSM (e.g. due to a reconfiguration) the hopping for the
TRX related to this hopping law is stopped in the cell, and the operator is informed.
Synthesizer frequency hopping is not affected.

Configuration of the BCCH Transceiver for EGPRS


The EGPRS PACKET CHANNEL REQUEST message, specific for EDGE mobile
stations (see "4.4.1 Packet Broadcast Control Channel (PBCCH)" and "9.8.2.4 TBF
Establishment for EDGE Mobile Stations"), is only supported by EDGE-CUs. Therefore
the use of the EGPRS PACKET CHANNEL REQUEST message can be forced by
configuring the BCCH-TRX in EDGE mode, which implies the allocation of a EDGE-CU
to the TRX, using the TRXMD parameter.
The BCCH carrier is continuously transmitted on all timeslots and without variation of RF
level; however, the RF power level may be ramped down between timeslots to facilitate
switching between RF transmitters. If 8PSK modulation is used for timeslots of the
BCCH TRX, these timeslots may use an 8PSK mean power which is at most 4 dB lower
than the power used for GMSK modulated timeslots. So, the use of the 8PSK modula-
tion on the BCCH carrier is critical and the operator can enable/disable it.
By means of the EBCCHTRX attribute, the operator can decide whether the channels
of the BCCH-TRX are available for EDGE 8PSK services or not. So, if the BCCH TRX
is enabled to support EDGE, but EBCCHTRX is set to FALSE, it means that only GMSK
modulated coding schemes of EDGE will be supported on the BCCH TRX (besides
GPRS coding schemes).
Even if the BCCH TRX is created in EDGE mode to support 8PSK modulation, the
i timeslot 0 is not allowed to use this modulation. This is necessary for compatibility with
Mobiles which do not support EDGE, in fact:
- the timeslot 0 is used to trasmit system information and signalling.

5.2 Configuration of GPRS Channels in a Cell


After having defined how many TRXs will support GPRS and EGPRS services (see 5.1),
the user indicates how the slots belonging to these TRXs will be managed; the following
statements show how the user can configure the GPRS/EGPRS service on the TRXs
where GSUP =TRUE:
1. the user can reserve, on a channel basis, some slots to packet switched services
only; these slots will be statically allocated to GPRS/EGPRS signalling (PBCCH
and PCCCH) and will not be used for circuit switched services. The user can define
these “only GPRS/EGPRS slots” on a channel basis, by setting the GDCH attribute
of the chosen CHAN object;
The GDCH attribute can assume the following values:
– PBCCH: i.e., the related channel is reserved for packet switched services, and
supports GPRS/EGPRS system information, common signalling, and data;

82 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

– PCCCH: i.e., the related channel is reserved for packet switched services, and
supports GPRS/EGPRS common signalling, and data.
Only one physical channel can be configured to carry the PBCCH logical
i channel (i.e., only one channel can be configured as the PBCCH); if the
operator then needs more PCCCHs, he must configure another channel as
PCCCH.
PBCCH and PCCCH channels can be defined on BCCH TRX only.

2. the user can then configure, among the remaining timeslots, other static timeslots
for PS services (i.e., not shared with CS services); the user can indicate this number
of static GPRS/EGPRS slots using the GMANPRES attribute of the PTPPKF object.
The difference, with respect to the configuration of static slots using the GDCH
attribute, is that with GDCH the configuration is made on a channel basis and
regards GPRS/EGPRS signalling channels only, whereas using GMANPRES the
configuration is made without indicating the channel, but only a “number of chan-
nels”, and regards GPRS/EGPRS traffic channels only.
If for example, the user defines 4 static slots for packet switched services, using
GMANPRES, then 4 slots will be reserved by the system for GPRS/EGPRS traffic
on the TRXs where GSUP =TRUE;
3. the user can choose among the remaining available slots (on TRXs where GPRS is
supported) the maximum number of dynamic GPRS/EGPRS channels; these chan-
nels will be shared between PS and CS services, according to the actual request of
resources.
To configure this number of shared slots, the user sets the GPDPDTCHA attribute
(PTPPKF object).
It indicates a percentage; this percentage is applied to the total number of available
slots (on TRXs where GPRS/EGPRS are supported) decreased by the number of
both static GPRS/EGPRS slots and slots reserved for GSM signalling. The
percentage indicates the maximum number of dynamic GPRS/EGPRS slots.
As it has been described, setting GSUP =TRUE means that that TRX is available for PS
i services in general; then, according to the TRXMD parameter value and to the avail-
ability of EDGE-CUs, TRXs are enabled to support GPRS only or both GPRS and
EGPRS.
So, if e.g. the user configure some static timeslots, according to the TRX where they are
assigned they will be used for GPRS only or for EGPRS too (obviously only on TRXs
with GSUP =TRUE).

To clarify previous concepts, let’s suppose (see Fig. 5.3) that three TRXs of a cell are
configured and enabled to support PS services (according to what has been described
in paragraph 5.1), and particularly:
– TRX0 and TRX4 support GPRS and EGPRS;
– TRX3 supports GPRS only.
Besides, the first TRX (TRX0) is the BCCH one and contains one SDCCH timeslot; the
second and the third TRXs (TRX1 and TRX2) are completely dedicated to circuit
switched services .
Then, on TRXs where packet switched services are supported, the total number of avail-
able slots for PS and CS services is equal to 22, in fact:
– 6 slots are available on TRX0;
– 8 slots are available on TRX3;
– 8 slots are available on TRX4.

A30808-X3247-L24-5-7618 83
GPRS/EGPRS Global Description Information
System

GDCH=PBCCH
TDMA frame
GSUP=TRUE
TRX 0 BCCH SDCCH TRXMD=EDGE
TRXCapability=EDGE
0 7
TDMA frame

TRX 1 SDCCH GSUP=FALSE

0 7
TDMA frame

TRX 2 SDCCH GSUP=FALSE

0 7
TDMA frame
GSUP=TRUE
TRX 3 TRXMD=GSM
TRXCapability=GSM
0 7

TDMA frame
GSUP=TRUE
TRX 4 TRXMD=EDGE
TRXCapability=EDGE
0 7

Fig. 5.3 Example of GPRS/EGPRS configuration.

If the user sets:


– GPDPDTCHA= 50 (i.e., 50%);
– GDCH= PBCCH for one CHAN object (CHAN:6) of the BCCH TRX (see Fig. 5.3);
– GMANPRES=1

then, the maximum number (N) of GPRS/EGPRS channels shared with CS services is
obtained by the following formula:

N= [Total number of available timeslots on TRXs with GSUP=TRUE - Number of


GPRS/EGPRS static slots(defined by GDCH and GMANPRES)]* GPDPDTCHA/100 =

= ( 22 - 2 ) * 50/100 = 10

So, in this cell we will have (on TRXs where GSUP=TRUE):


– 2 slots statically allocated for packet switched services (one signalling slot defined
on a channel basis using the GDCH attribute and the other one defined on a cell
basis using GMANPRES)
– 10 slots shared between PS and CS services (according to the previous formula and
the GPDPDTCHA setting);

84 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

– 10 slots reserved for CS services (i.e., the remaining slots on the TRXs where
GSUP=TRUE).
Obviously both the TRX1 and the TRX2 will be used for circuit switched services only.
It must be underlined that the previous example is only valid if HOPMODE=SYNHOP.
i As soon as HOPMODE=BBHOP (independently of whether or not frequency hopping is
enabled), timeslots 0 of all non-BCCH TRX are never allocated for GPRS/EGPRS.

Finally, to define how many users can be multiplexed in a PDCH, the GMANMSAL
attribute (PTPPKF object) is used. It defines the maximum number of GPRS/EGPRS
users that can share the same timeslot (PDCH); it is composed of two fields: the first
indicates the maximum number of users in the uplink direction, the second one specifies
the maximum number of users in the downlink direction.

5.3 Management of Packet Data Channels

5.3.1 Generalities about Resource Assignments


The radio resource assignment to a mobile station requiring an uplink or a downlink TBF
establishment is performed under the BSC control, through co-operation between the
PPCU/PPXU processors (involved in PDCH management and scheduling functions for
UL and DL data transfer, see "6 Hardware and Software Architecture") and the TDPC
processor (dedicated to resource management).
When a new GPRS/EGPRS request arrives at the BSC, the PPCU/PPXU, starting from
the MS multislot class and the required peak throughput, estimates the number of radio
resources (i.e., the number of PDCHs) requested for TBF establishment.
In order to provide the best possible throughput for each user, the system considers the
Peak Throughput Class to calculate the number of resources to be assigned to a new
TBF. The resource allocation algorithm tries to assign to each TBF a number of timeslots
that depends on the required Peak Throughput Class, i.e.;
a) MS Multislot Class (it is sent from the MS to the network during the GPRS attach
procedure, see "9.3.2.1 Attach Function");
b) Peak Throughput:
– when uplink TBF is derived from the Channel Request description inside the
Packet Resource Request or Packet DL Ack/Nack;
– when downlink TBF is derived from the Qos Profile IE of the DL-UNITDATA;
c) Coding Scheme to apply, given in throughput per timeslot.
Trying to maximize the throughput is the most important criteria in radio resource
i search.

The PCU, after calculating the number of requested resources, checks if the actual used
strategy is the vertical one or the horizontal one (see "5.3.2 Horizontal/Vertical Alloca-
tion Strategies"). Then, according to the used strategy and to the needed resources, it
sends the correct request to the TDPC.
According to the requests received by PPCU/PPXU, the TDPC is responsible for:
1. the assignment of the proper radio resources on the air interface (PDCHs);
2. the assignment of the Abis interface subslots related to these PDCHs.
When the PCU request arrives at the TDPC, the TDPC tries to satisfy the request.

A30808-X3247-L24-5-7618 85
GPRS/EGPRS Global Description Information
System

If required channels are found, the TDPC sends the ACK message to the PCU, other-
wise other actions have to be executed (see "5.3.3.2 TDPC Algorithm").
Note that, on a cell basis, the PPCU/PPXU knows:
1. the number of PDCHs in use at a given time; i.e., it knows:
– the timeslots (PDCHs) with at least one TBF assigned;
– PDCHs for which the Empty Channel Timer is running.
In fact, when the last MS associated to a PDCH is released, the “virtual” assignment
persists for the duration of the Empty Channel Timer. The value of this timer is set,
by means of the TEMPCH parameter, to avoid continuous requests (in cases of high
GPRS/EGPRS traffic) from the PPCU to the TDPC.
When the timer is still active, the allocated PDCH(s) for the “released” TBF are still
seen as allocated even if they are no longer active.
2. the number of PTDs (equal to the number of Abis subslots) related to the PDCHs
still in use.
From this point, it could happen that (according to what has been described in
"6.3.1 Concatenated PCU Frames") for a PDCH one or more corresponding PDTs
are useless, i.e., they are filled with idle PCU frames, due to downgrade to a coding
scheme needing less PDTs than the initial ones. When a PDT is filled with idle PCU
frames, the PCU, before releasing it waits until a timer defined by the TEMPPDT
parameter expires.
The timer is used to avoid continuous requests of Abis resources from the PPCU to
the TDPC; in fact to every PDT has corresonding Abis 16 kbit/s subslots that the
PCU requires for the TDPC, since it is the TDPC that manages Abis resources.

5.3.2 Horizontal/Vertical Allocation Strategies


When a new request of GPRS/EGPRS resources arrives at the BSC, two different strat-
egies are provided to assign packet switched data channels; they are:
• the VERTICAL ALLOCATION (VA) strategy: before assigning a new slot to
GPRS/EGPRS service, the already used slots must be filled as much as it is
possible, according to the chosen GMANMSAL value (see 5.3.2.1);
• the HORIZONTAL ALLOCATION (HA) strategy: it is introduced in the system to
allow higher bit data rates when the cell is not congested.
This strategy is intended to distribute the incoming GPRS/EGPRS calls on all the
available PDCHs. In this way not too many users are multiplexed on the same
PDCH, increasing the data transfer throughput for all the involved mobiles
(see 5.3.2.2).
The user can manage the allocation strategy according to his needs, by means of
specific parameter settings. It is important to underline how the chosen strategy
depends upon both from radio resources availability and Abis resources availability.
Chapter 5.3.2.3 describes how the radio interface situation triggers the switching from
HA to VA and vice versa, according to parameter settings; whereas chapter 5.3.2.4
describes the analogous topics for what concerns the Abis interface. Finally, the
complete algorithm is summarized in chapter 5.3.2.5.

5.3.2.1 Vertical Allocation Strategy (VA)


When GPRS/EGPRS channels are handled using the Vertical Allocation (VA) algorithm,
the criterion is to multiplex the maximum number of mobiles on each channel, before

86 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

assigning a new PDCH. This is obtained by filling an already used PDCH, as much as
it is possible, compatibly with:
• network settings of GMANPRES, GPDPDTCHA and GMANMSAL attributes;
• MS multislot capability.
If, for example, 2 mobile stations perform 3DL+1UL GPRS/EGPRS calls, the BSC e.g.,
will assign them the timeslots number 2, 3 and 4.
In this way, timeslots from 5 to 7 remain free because the BSC multiplexes the 2 mobiles
on the same 3 PDCHs, as drawn in Fig. 5.4.

TS0 TS1 TS2 TS3 TS4 TS5 TS6 TS7

BCCH TRX bcch sdcch

MS1DL MS1DL MS1DL


MS1UL

MS2DL MS2DL MS2DL


MS2UL

Fig. 5.4 Example of Vertical Allocation Algorithm

When the vertical allocation strategy is used, the BSC tries to multiplex, in a fair way,
the mobile requests, using “flat distribution”. With flat distribution, if the BSC is in VA
condition and over each radio timeslot is multiplexed only one mobile station, if three
GPRS/EGPRS mobile requests (single slot) arrive to BSC, the BSC will multiplex the 3
mobiles over 3 different radio channels trying to uniformly distribute the resources.
If flat distribution was not used, all the 3 mobiles would be multiplexed in the same
timeslot (compatibly with GMANMSAL setting).

5.3.2.2 Horizontal Allocation Strategy (HA)


The Horizontal Allocation (HA) strategy is intended to distribute the incoming
GPRS/EGPRS calls on all the available PDCHs. In this way not too many MSs are multi-
plexed on the same PDCH increasing the data transfer throughput for all the involved
mobiles.
When a new request of PDCH channels arrives at the BSC and radio channels for
GPRS service are still available, the BSC assigns new radio channels to the
GPRS/EGPRS mobiles instead of increasing the number of mobiles multiplexed on the
already busy channels.
If, for example, 2 mobiles perform 3DL+1UL GPRS/EGPRS calls, the BSC will assign
timeslots number 2, 3 and 4 to the first call, then timeslots number 5, 6 and 7 to the
second call.
In this way each timeslot is used for a lower number of calls and the throughput is better
than that for the vertical allocation strategy, as drawn in Fig. 5.5.

A30808-X3247-L24-5-7618 87
GPRS/EGPRS Global Description Information
System

TS0 TS1 TS2 TS3 TS4 TS5 TS6 TS7

BCCH TRX bcch sdcch

MS1DL MS1DL MS1DL MS2DL MS2DL MS2DL


MS1UL MS2UL

Fig. 5.5 Example of Horizontal Allocation Algorithm

5.3.2.3 Switching between VA and HA According to Radio Conditions


Taking into account the radio interface, the BSC switches autonomously from VA to HA
(and vice versa) in relation to the traffic load in the cell (i.e., in relation to the number of
busy air timeslots).
The aim of this feature is to use the horizontal allocation when the cell is not loaded;
otherwise the adopted strategy will be the vertical one.
To avoid frequent changes between HA and VA strategies, two thresholds are defined:
– one threshold (ThresholdIdleChannelHV) represents the transaction from the hori-
zontal allocation to the vertical one;
– the other threshold (ThresholdIdleChannelVH) represents the transaction from the
vertical allocation to the horizontal one.
These thresholds, used to activate horizontal/vertical allocation, are managed by the
GASTRTH attribute (PTPPKF object).
The GASTRTH contains a third field, called ThresholdIdleChanEU; this field represents
i a threshold that is used in the radio resource upgrade strategy ("5.3.4.1 Upgrade of
Radio Resources").

The threshold, that causes the transaction from one allocation algorithm to the other
one, represents the percentage of idle slots in the whole cell.
The percentage is calculated as the number of idle channels in the cell with respect to
the number of available channels in the cell (TCHs or PDCHs; do not consider slots
containing GSM signalling, such as BCCH or SDCCHs slots, and also slots statically
reserved to GPRS/EGPRS).
The number of available channels in the cell is calculated as:
Available Channels = Total number of configured channels - Number of OUT OF
SERVICE channels - Number of GPRS/EGPRS static channels (defined by both GDCH
and GMANPRES) - Number of GSM signalling channels

Obviously the number of Idle Channels is the number of “not busy” channels inside the
pool of all the available channels of the cell.
Then the percentage of idle channels in the cell (to be compared with the thresholds of
the GASTRTH parameter) is given by:

Percentage of Idle Channels in the cell = Idle Channels / Available Channels

88 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

It is important to highlight that the evaluated value, that represents the percentage of
i idle channels in the cell, is truncated, so decimals are not taken into account in the
comparison with thresholds. For example, if the internal evaluation estimates that the
percentage of idle channels in the cell is 10.9%, then the real value that is compared to
the thresholds is 10% and not 11%).

Let’s consider a cell with five configured TRXs, three of them supporting GPRS/EGPRS
(see Fig. 5.6) where:
– TRX0 contains BCCH and SDCCH logical channels;
– TRX1 and TRX2 do not support GPRS/EGPRS;
– TRX3 supports the GPRS/EGPRS service, but it is out of service.
– the timeslots of the TRXs (with the exception of the BCCH and the SDCCH ones)
are defined as TCHF_HLF; then each timeslot represents, from the circuit switched
services point of view, two available channels.

TDMA frame

TRX 0 BCCH SDCCH GSUP=TRUE

0 7
TDMA frame

TRX 1 SDCCH GSUP=FALSE

0 7
TDMA frame

TRX 2 SDCCH GSUP=FALSE

0 7
TDMA frame

TRX 3 GSUP=TRUE

0 7

TDMA frame

TRX 4 GSUP=TRUE

0 7

Fig. 5.6 Example of a Cell Configured with Five TRXs.

A30808-X3247-L24-5-7618 89
GPRS/EGPRS Global Description Information
System

In this case:
• total number of configured traffic channels = (6 + 7 + 7 + 8 + 8 ) * 2 = 72;
BCCH and SDCCH signalling channels are not considered, only traffic chan-
i nels are taken into account.

• number of OUT OF SERVICE channels = (8 * 2) = 16;


• if we set GMANPRES= 3 (without setting the GDCH value for any channel), then the
Number of static GPRS/EGPRS channels to be considered in the previous formula
is equal to 6 (since each timeslot reserved to packet switched services represents,
from the circuit switched services point of view, two available channels).
Then, according to the above formula (without considering GSM signalling channels):
Available channels = 72 - 16 - 6 = 50.

If, for instance, the GASTRTH parameter has been set with the following values:
– ThresholdIdleChannelHV=30%, for the transaction from HA ---> VA;
– ThresholdIdleChannelVH=40%, for the transaction from VA ---> HA.

when the percentage of idle slots is over 40%, horizontal allocation is used. In this case:
(100 * Idle channels) / Available channels > 40 ----> Idle channels = 21.
So, when in the cell the number of idle channels equals 21, the Horizontal Allocation
strategy is used.

If the percentage of idle slots falls under 30%, vertical allocation is used; in this case:
(100 * Idle channels) / Available channels < 30 ----> Idle channels = 14.
So, when in the cell the number of idle channels reaches 14, the vertical Allocation
strategy is used.
If the percentage again exceeds the 40% threshold, the horizontal allocation algorithm
is restored.
The difference between the two thresholds of the GASTRTH parameter should not be
i too high, but the thresholds have to be set to reasonable values (also taking into account
the number of configured TCHs in the cell). Otherwise it could happen that, when
VERTICAL allocation is used, a return back to HORIZONTAL one is applied only when
the cell is completely idle, and this is not a real hysteresis behavior.

It should be noted that when horizontal allocation is used to assign GPRS/EGPRS


i resources, two conditions, related to radio interface, can determine the transition to the
vertical one: the first condition occurs when, in the cell, the number of idle channels falls
below the threshold set by the GASTRTH parameter; the second one occurs when the
number of channels assigned to GPRS/EGPRS users reaches the maximum number of
channels configured for PS services by means of the GMANPRES, GDCH and GPDP-
DTCHA parameters.

90 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

5.3.2.4 Switching between VA and HA according to Abis Interface Condi-


tions
Taking into account the Abis interface, the BSC switches autonomously from VA to HA
(and vice versa) in relation to the number of exploited Abis resources (i.e., in relation to
the number of busy Abis subslots).
In fact, according to the flexible Abis allocation strategy (see 6.3), the number of used
Abis resources of a BTSM depends on which services are currently running on the cells
managed by the BTSM. Then it is very important to check, besides the radio interface
situation, the level of congestion of the Abis interface, to take the right countermeasures.
So the switching from vertical allocation to horizontal allocation and vice versa is also
influenced from the Abis interface.
To manage the resource allocation process, a set of thresholds regarding Abis alloca-
tion usage is introduced. The user can set these thresholds by the GASTRABISTH
parameter (gprsAllocationStrategyAbisThresholds) of the BTSM object.
GASTRABISTH is composed of four fields, two of them allow management of the
vertical/horizontal allocation strategy and are here described; the remaining two regard
the upgrade of Abis resources and are discussed in "5.3.4.2 Upgrade of Abis
Resources":
• the first threshold (thresholdIdleAbisHV) defines the percentage of idle Abis subslots
of the BTSM (over the available Abis subslots of the BTSM) under which vertical
allocation strategy for Abis scarcity is activated on the radio interface (for all the cell
of the BTSM). Activating the vertical allocation in cases of Abis scarcity is useful to
re-use in multiplexing the already allocated Abis subslots, slowing down the alloca-
tion of new Abis resources;
• the second one (thresholdIdleAbisVH) defines the percentage of idle Abis subslots
of the BTSM (over the available Abis subslots of the BTSM) over which horizontal
allocation can be activated again on the radio interface, if the thresholds on radio
resources (on a cell basis) also allow that.
Constraints on these two Abis thresholds are:

thresholdIdleAbisHV < thresholdIdleAbisVH

5.3.2.5 Allocation of Resources


Besides the situations described in 5.3.2.3 and 5.3.2.4, in the BR7.0 release, the
switching to vertical/horizontal allocation is also driven by the availability/unavailability
of PDTs on PCU (remember that with the high capacity BSC, see "6.1 Supported BSC
Types", the maximum number of GPRS channels manageable by the single PCU, i.e.,
the maximum number of PDT manageable by the single PCU, is fixed to 256).
Then according to the availability/unavailability of PDTs on the PCU side:
– when the percentage of busy PDTs on a PCU is 100%, then vertical allocation is
applied;
– when the percentage of busy PDTs on a PCU is less than100%, then horizontal allo-
cation is applied (provided that thresholds of GASTRTH and GASTRABISTH
parameters do not lead to vertical allocation).
In summary, in cells belonging to a BTSM with dynamic Abis management, the following
situations are possible during the allocation of radio resources, according to different
contexts:

A30808-X3247-L24-5-7618 91
GPRS/EGPRS Global Description Information
System

a) horizontal allocation in a cell is used if at the same time these three conditions are
satisfied:
– there is no radio scarcity in the cell, i.e., the percentage of idle air timeslots in the
cell is greater than the ThresholdIdleChannelHV field of the GASTRTH param-
eter;
– there is no Abis resources scarcity, i.e., the percentage of idle Abis subslots of the
BTSM managing the cell is greater than the thresholdIdleAbisHV field of the
GASTRABISTH parameter;
– there is no PDT exhaustion for the PCU that manages the cell.
b) starting from horizontal allocation, if there is radio scarcity, i.e., the percentage of
idle air timeslots in the cell becomes lower than the ThresholdIdleChannelHV field
of the GASTRTH parameter, than vertical allocation is triggered.
c) starting from horizontal allocation, if there is Abis scarcity, i.e., the percentage of idle
Abis subslots in the BTSM becomes lower than the ThresholdIdleAbisHV field of the
GASTRABISTH parameter, than vertical allocation is triggered; the PCUs that are
handling cells belonging to the impacted BTSM are informed.
d) starting from horizontal allocation, if there is PDT exhaustion in the PCU, than
vertical allocation is triggered.
e) if the vertical allocation of the cell is due to radio scarcity only, and the percentage
of idle air timeslots in the cell becomes greater than the ThresholdIdleChannelVH
field of the GASTRTH parameter, than horizontal allocation is triggered.
f) if the vertical allocation of the cell is due to Abis scarcity only, and the percentage of
idle Abis subslots in the BTSM becomes greater than the ThresholdIdleAbisVH field
of the GASTRABISTH parameter, than horizontal allocation is triggered; the PCUs
that are assigned cells belonging to the impacted BTSM are informed.
The allocation strategy used is managed and implemented in the TDPC. The TDPC
informs the PCU about the used strategy via the allocation flag (HA/VA). This flag is
updated each time the TDPC replies to PCU requests for resources.
To avoid possible misalignment between TDPC and PCU, as regards the allocation flag,
a mechanism is foreseen for which an audit, running every 10 seconds for each
equipped and in service PCU, is sent to communicate to the PCU the current allocation
strategy used on the TDPC side.

5.3.3 Management of Incoming GPRS/EGPRS Requests


With the introduction of packet switched data calls, a new clear and flexible strategy for
channel allocation is required. The introduction of channel allocation strategies requires
a mechanism to use/reuse the system resources, in order to better comply with the oper-
ator’s choices. In this sense it is very important to optimize the resource allocation in
order to use each TRX in the best possible way, and to satisfy each new request
according to the customer’s setting (both for Data and for Speech calls).
Regarding the management of different services, the operator can configure in the cell
static and dynamic GPRS/EGPRS timeslots (see "5.2 Configuration of GPRS Channels
in a Cell").
When a new GPRS/EGPRS request arrives at the BSC, it starts a process that is
responsible for the management of all the incoming requests: the PCU, when it is
needed, asks to the TDPC to set-up new channels for packet switched services. The aim
of the task in PCU/TDPC is to allocate the requested resources according to the oper-
ator setting and to the Mobile Station preferences.

92 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

The radio resources allocation algorithm keeps into account the availability of EGPRS
service, and the presence of EDGE capable mobiles and TRXs: in fact MSs that support
EGPRS could be assigned either to TRX supporting EDGE (exploiting EGPRS coding
schemes) or to TRX supporting GPRS (in this case, only GPRS coding schemes can be
used).
To distinguish EDGE TRXs from non-EDGE TRXs, both on TDPC and PCU, the
Resource Manager looks at the ‘TRX availability’ dynamic attribute of the TRX (see
"5.1.3 Aspects Related to Carrier Configuration"): TRX with ‘unknown’ ‘TRX availability’
are not taken into account, since they are not available for service at all.
Both on TDPC and on PCCU, the radio resource research algorithm will also take into
account the EBCCHTRX attribute, that specifies whether the 8PSK modulation is
allowed on the BCCH TRX (see " Configuration of the BCCH Transceiver for EGPRS").
The aim of the algorithm is: maximize the throughput in the limits of specified peak
throughput (if specified), minimizing the number of allocated radio resources.
Therefore, in principle, EDGE TRXs are preferable for EDGE-capable mobiles, because
higher data rates are possible with a lower number of radio resources, and even when
data rates are comparable, or GPRS data rates are slightly better (e.g., CS4 versus
MCS4), we can expect better performances from a TBF operating in EGPRS mode
(instead of GPRS mode), due to specific retransmissions rules and incremental redun-
dancy (see "9.9.1.2 EGPRS Acknowledged Mode").
But, for an EGPRS TBF, when the multislot capability of a mobile is high, and if radio
resources are insufficient on EDGE TRXs and available on non-EDGE TRXs, a non-
EDGE TRX could be preferable. In fact, as it has been said, trying to maximize the
throughput is the most important criteria in radio resource research. Let us consider an
example; a request to establish a TBF with the following requirements arrives at the
BSC:

Peak throughput = 80 kbit/s


Multi-slot capability = 4 timeslots

According to the request, the BSC finds two solutions; the first using NE timeslots on an
EDGE TRX, the second using NG timeslots on a non EDGE TRX, where:

NE = 2 (with MCS9)
NG = 4 (with CS4)

The ‘best’ solution is to allocate 2 radio timeslots on the EDGE TRX, because the peak
throughput is sustained with the minimum number of radio resources.
But if only 1 radio resource (NE’) is available on an EDGE TRX, the sustainable
throughput is only 59.2 kbit/s. In this case, 3 radio resources (62.4 kbit/s), or 4 radio
resources (83.2 kbit/s), on a non-EDGE TRX allow sustaining of the peak throughput,
and should be considered better solutions.
From the configuration point of view, to allow, both for the voice and data calls, an higher
flexibility for different operator’s strategies, a parameter is provided. This parameter,
called CPOLICY, allows the operator to indicate on which TRX (BCCH or not BCCH) a
certain type of call (voice or data) will be preferably allocated. In this way, a clear usage
policy for the BCCH TRX channel allocation is guaranteed. The PCU and the TDPC

A30808-X3247-L24-5-7618 93
GPRS/EGPRS Global Description Information
System

administer this setting when they have to assign resources to GPRS users (see
"5.3.3.1 PCU Algorithm" and "5.3.3.2 TDPC Algorithm").
So, taking account GPRS and EGPRS mobiles, TRXs supporting EGPRS or GPRS
only, and the CPOLICY parameter, the research algorithm basically follows these rules :
• in cases of EDGE capable mobiles, TRXs are sorted giving priority to the EDGE
TRXs; this criterion is more important than the call policy. That is: if the CPOLICY
parameter is set to DATA_CALL_ON_BCCH and the BCCH TRX is a non-EDGE
TRX, the BCCH TRX is checked AFTER all the EDGE TRXs and before all the other
non-EDGE TRXs.
When the EDGE BCCH TRX doesn’t support 8PSK (EBCCHTRX=FALSE), the Call
Policy is disregarded: BCCH TRX will be considered after all the other EDGE TRXs
even if the CPOLICY parameter is set to DATA_CALL_ON_BCCH. If the operator
wants to give more priority to the BCCH TRX, according to the Call Policy, the
EBCCHTRX attribute should be set to TRUE;
• in cases of non-EDGE capable mobiles, TRXs are sorted giving priority to the non
EDGE TRXs; this criterion is more important than the call policy. That is: if the
CPOLICY parameter is set to DATA_CALL_ON_BCCH and the BCCH TRX is a
EDGE TRX, the BCCH TRX is checked AFTER all the non-EDGE TRXs and before
all the other EDGE TRXs.
The horizontal/vertical allocation algorithm on TDPC receives as input a
PDCH_Request message from the PCU containing, among other information, a list of
suggestions for channels to be granted by TDPC. The “already busy for GPRS/EGPRS”
channels can be assigned only by the PCU, while idle channels can be assigned only
by the TDPC.
If the incoming GPRS request cannot be satisfied (because some timeslots have to be
free for the GPRS multislot calls, or because the cell is congested), the request is
inserted in a waiting queue (i.e., a ‘stand by’ queue), and it will be served as soon as the
proper actions have been performed (see "5.3.6 Waiting Queue Management").
The waiting queue where the “not served GPRS requests” are inserted, is different
! from the queue related to the Queuing feature. In fact the Queuing feature is related to
The

circuit switched calls only, and the related queue is called queuing list.

As it has been described, the algorithm used to assign GPRS resources is split in two
parts: one is performed on the PCU and the other one on the TDPC; in the following
sections the parts are described.

5.3.3.1 PCU Algorithm


When a new request of GPRS resources arrives at the BSC, the first actions are taken
by the PCU that handles the cell from which the request is arrived.
This chapter describes the PCU algorithm in cases of new GPRS/EGPRS calls. The
i upgrading procedure, in cases of new resources requested for already established
calls, are discussed in "5.3.4 Upgrading Strategies".

When a new request is sent to the PCU, the following information is provided:
• Mobile capability (GPRS or EGPRS);
• required Peak throughput;
• Multislot class;
• Candidate Initial Coding Scheme (CS/MCS); as it has been described in
"4.2 Channel Coding", the user can set the preferred initial coding scheme, for both

94 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

GPRS and EGPRS services, to be used when a new TBF starts. The O&M config-
ured initial coding schemes are only used if no information about a MS in a cell is
available when a new TBF starts. In fact, the PCU holds in memory, for each mobile
station, the last coding scheme (either CS or MCS) used in the uplink/downlink
directions for TBFs associated to the MS; neverthless the PCU maintains this infor-
mation only for a specific period of time. So the Candidate Initial Coding Scheme will
be:
– the coding scheme stored in the PCU memory, if this information is still available;
– otherwise the O&M configured value.
To get more detailed information about initial coding scheme handling,
i please refer to "10.5.3 Selection of the Candidate Initial Coding Scheme".

The aim of the search on the PCU side is to find a number of adjacent PDCHs in order
to maximize the throughput of the TBF.
The PCU, before starting to search radio resources on the TRXs, calculates the optimal
number (N) of radio resources that allow the maximum “initial target throughput” of the
data transmission.
The general formula to calculate the number of “optimal” number of radio resources (N)
is the following:

Peak Throughput Available Peak Throughput


NOT Available
N Min(ceil ( PT / (T_I_CS x (1-BLER) ), Multislot Class) Multislot Class

where:
ceil = round up to the upper integer
PT = required Peak Throughput
T_I_CS =throughput (maximum data rate) of “Candidate initial coding scheme”
BLER = it is the initial BLER value.
The BLER value is defined as the number of radio blocks to be repeated (not acknowl-
edged blocks) versus the number of transmitted radio blocks in total (i.e., the sum of
the acknowledged blocks and the not acknowledged one, see "9.9 RLC Data Block
Transfer"):

BLER= NACK_Blocks/(ACK_Blocks+NACK_Blocks)

The user can define the initial BLER value, used in the resource assignment process,
via the INIBLER parameter.
The O&M configured initial BLER is only used if no information about a MS in a cell is
available when a new TBF starts. In fact, the PCU stores in memory, for each mobile
station, besides the last coding scheme, the last measured BLER value (historical
BLER) associated to the MS; neverthless the PCU maintains this information only for
a specific period of time. The Initial BLER corresponds to the INIBLER value if no
“historical BLER” information is available; otherwise the “historical BLER” is used.

The optimal number of radio resources that the PCU calculates depends on:
– the availability of the Peak Throughput in the request;
– the mobile station capability, i.e., if the MS is EGPRS capable or not.
The different possibilities are described:

A30808-X3247-L24-5-7618 95
GPRS/EGPRS Global Description Information
System

a) In cases of mobiles with EGPRS capability and in cases where the peak throughput
is available, two calculations must be performed, for a ‘pure’ UL or DL TBF setup (no
concurrent TBF in progress):
– calculation for the ‘optimal’ number NE of radio resources for EDGE TRX (based
on the ‘candidate’ initial MCS);
– calculation for the ‘optimal’ number NG of radio resources for non EDGE TRX
(based on the ‘candidate’ initial CS);
– in cases where a concurrent TBF is in progress with TBF mode EDGE, only NE
will be calculated; in cases where a concurrent TBF is in progress with TBF mode
GPRS, only NG will be calculated; this is because, if a MS is assigned concurrent
TBFs, these will be in the same TBF mode.
b) In cases of mobiles without EGPRS capability and in cases where the peak
throughput is available, only the calculation for NG is performed;
c) When the peak throughput is not available, the multislot class is taken into account.
Then, different solutions (i.e., different radio timeslot configurations) are compared in
terms of ‘initial target throughput’ instead of ‘number of timeslot’; the basic formula to
calculate the initial target throughput per timeslot is:

Initial target throughput per timeslot = throughput (maximum data rate) of the candi-
date initial CS/MCS

This value is multiplied by the number (NG or NE) of radio resources to get the better
solution; the better solution is that which provides the highest Initial Target Throughput.
When the initial target throughput per PDCH on GPRS TRXs is slightly better than the
i initial target throughput per PDCH on EDGE TRXs, solutions allocating N radio
resources on EDGE TRXs are preferred to solutions allocating N radio resources on
GPRS TRXs, because better performances are expected from EGPRS specific retrans-
mission rules and incremental redundancy (see "9.9.1.2 EGPRS Acknowledged
Mode"). This situation can occur, for example, when the MCS and CS used to calculate
the ‘initial target throughput’ are ‘homologous’ (e.g., CS4/MCS4). For example, 3 radio
timeslots in EGPRS TBF mode are preferable to 3 radio timeslots in GPRS mode, in
case the initial MCS in the cell is MCS4 (data rate 17,6) and the initial GPRS CS in the
cell is CS4 (data rate 20,8).

The “Initial target throughput” is just an indicator, used to compare different radio
i timeslot configurations; there is no guarantee that the ‘initial target throughput’ is really
achieved, because the actual throughput depends on several factors: radio conditions,
C/I, Link Adaptation, multiplexing factor, availability of Abis and PDT resources, etc. In
particular, in cases of Abis/PDT resources scarcity it is not guaranteed that the resource
assignment will result in the best solution in terms of throughput (see "5.3.3.2 TDPC
Algorithm").

When the PCU has calculated the optimal number of radio resources, it starts executing
a pre-search of radio resources on available TRXs; a different process is applied
according to the allocation strategy currently in use (the PCU algorithm is shown in
Fig. 5.7).
In cases of Horizontal Allocation strategy the PCU starts a search on all the TRXs
usable for GPRS or EGPRS according to the kind of request. The criteria used to find
resources are the following (in order of priority):
1. prefer EGPRS on EDGE TRXs and GPRS on non-EDGE TRXs;
2. maximize the “Initial target throughput”;

96 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

3. maximize the number of empty channels (i.e., the channels already allocated in
packet transfer mode, but without assigned TBF; these channels are seen by the
PCU as allocated until the TEMPCH timer expires);
4. minimize the overall weight on the affected PDCHs;
The following QoS parameters are taken into account in the resource allo-
i cation process on PCU side:
- Radio Priority in the uplink direction;
- Service Precedence in the downlink direction.
Internally, UL radio priority and DL service precedence are mapped into a
unique ‘internal priority’ so that UL and DL TBFs are comparable. Internal
priority’ here mentioned coincides with the ‘scheduling priority’ used by the
scheduler process (see "9.9.7 Notes About GPRS/EGPRS TBF Sched-
uling" to read how Qos attributes are mapped to scheduling priority).
According to its priority, each TBF is assigned a ‘weight’; as described in
9.9.7, the association between priorities and weights is performed by the
following O&M attributes: SCHWEIPRI1, SCHWEIPRI2, SCHWEIPRI3,
SCHWEIPRI4.
So, each PDCH(i) is assigned a ‘total weight’ W(i) given by:
W(i) = Sum of W(k)
where W(k) is the weight of all the TBF(k) multiplexed on the PDCH(i).
On PCU, the algorithm for radio resources presearch, in addition to the
other criteria, tries then to minimize the total weight of the suggestions to
be sent to TDPC.

5. maximize the number of adjacent timeslots with respect to the ones already in
packet transfer mode;
6. choose the preferred TRXs according to the CPOLICY parameter.
The output of this algorithm is a possible configuration on one TRX. Two cases exist:
1. if all the chosen timeslots are already available at PCU side, i.e., the PCU does not
need to ask new idle PDCH channels to the TDPC, the timeslots are assigned by
PCU immediately (i.e., no PDCH_Request message is sent to TDPC). But in this
case, according to the flexible Abis allocation strategy, it could happen that, even if
no new PDCH has to be allocated, new PDT/Abis allocation is necessary to support
the new TBF; this is because e.g., the current Abis allocation is not enough to
support the candidate initial coding scheme. In this case, the PCU will sent a request
to the TDPC for additional Abis resources using the PDCH_Abis_Upgrade message
(see "5.3.4.2 Upgrade of Abis Resources").
It must be noted that when horizontal allocation is used, the timeslots already avail-
able at PCU side are the empty channels, i.e., free PDCHs for which the TEMPCH
timer is running (these channels are also called pre-allocated)
2. in case some timeslots are not immediately available, i.e., when new idle channels
are necessary at the PCU side, a PDCH_request message is sent to the TDPC indi-
cating this configuration as a suggestion (the request also notifies the TDPC of the
“Initial Target Throughput per timeslot). The request also contains the number of
Abis resources needed to support the TBF.
In order to handle parallel requests, the TRX belonging to this suggestion is set as
“frozen” and excluded from subsequent searches until either the TDPC answers
(positively or not) or a protection timer expires.
With the Vertical Allocation strategy, the idea is to reduce the number of new timeslots
to asked of the TDPC for the incoming request.

A30808-X3247-L24-5-7618 97
GPRS/EGPRS Global Description Information
System

When the Vertical Allocation strategy is used, the layering method is the following (flat
distribution): instead of multiplexing continuously on the same timeslot (until the
GMANMSAL value is reached), the TBFs are spread on all the already assigned
timeslots, on all the TRXs. This leads to better system performances in terms of TBF
throughput. This is done by multiplexing the new TBF on the timeslots already in packet
mode that are not in the busy state (the busy state is set when the number of TBFs multi-
plexed on a PDCH reaches the GMANMSAL value).
The criteria used to find resources are the following (in order of priority):
1. prefer EGPRS on EDGE TRXs and GPRS on non EDGE TRXs;
2. maximize the initial target throughput;
3. maximize the number of empty channels;
4. minimize the overall weight on the affected PDCHs;
5. choose the preferred TRXs according to the CPOLICY parameter.
The output of this algorithm is a possible configuration on one TRX. If all the chosen
timeslots are already available at the PCU side, they are assigned immediately.
It must be noted, that when vertical allocation is used, timeslots already available at PCU
side are:
– timeslots already assigned to GPRS users, containing active TBFs;
– empty channels, i.e., free PDCHs for which the TEMPCH timer is running (these
channels are also called pre-allocated).
In this case no new PDCH has to be allocated, but it could happen that the current
PDT/Abis allocation is not enough, so the PCU could send a request to TDPC for addi-
tional Abis resources by the PDCH_Abis_Upgrade message (to have more details about
upgrade of Abis resources, see "5.3.4.2 Upgrade of Abis Resources").
In case some timeslots are not immediately available, a PDCH_request message is sent
to the TDPC indicating the suggestion to be preferred in the search. Also in this case, in
order to handle parallel requests, the TRX belonging to the suggestions is set as
“frozen” and excluded from subsequent searches until either TDPC answers (positively
or not) or a protection timer expires.

98 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

MS/SGSN
request

Calculate the number of


requested PDCHs according to:
- Multislot Class
- Peak Throughput
-Mobile Station Capability
-Candidate Initial Coding Scheme

Horizontal Allocation Vertical


type?

Find channels Find channels


according to HA according to VA
criteria criteria

PCU PCU
needs new YES NO needs new
NO YES
PDCH PDCH
channels? channels?

Send PDCH_Request Send PDCH_Request


to TDPC to TDPC

PCU PCU
needs new YES needs new YES
PDT? PDT?

NO NO

Assign channels Send Assign channels Send


already in PDCH_Abis_Upgrade already in PDCH_Abis_Upgrade
Packet Mode to TDPC Packet Mode to TDPC

Fig. 5.7 Allocation Algorithm followed by the PCU

A30808-X3247-L24-5-7618 99
GPRS/EGPRS Global Description Information
System

5.3.3.2 TDPC Algorithm


As it has been described (see 5.3.3.1), the PCU can send to the TDPC two kinds of
message:
a) the PDCH_Request message, when new PDCHs on the air interface must be
assigned; this message also specifies the number of Abis resources needed to
support the TBF. New PDCHs can be required in the cases that follow:
– when a new PS call must be established;
– when a PS call already established needs more radio resources (i.e., in cases of
upgrade of radio resources);
b) a PDCH_Abis_Upgrade message, when new PDCHs on the air interface must not
be assigned, but new Abis resources are needed. This message is sent in the
following cases:
– when a new TBF, which uses a coding scheme requiring a certain number of Abis
resources, is allocated with the vertical allocation on PDCHs already filled with
TBF requiring less Abis resources;
– when the upgrade of the coding scheme of one TBF requires that more Abis
resources are assigned.
The cases requiring the upgrade of already assigned resources (PDCHs or Abis) are
discussed in "5.3.4 Upgrading Strategies"; the case regarding new incoming PS calls is
here described.
When the PCU needs new PDCH channels to be assigned to the incoming
GPRS/EGPRS call, it sends to the TDPC a PDCH_Request message.
The PDCH_Request message received from the PCU contains the following informa-
tion:
• Number of timeslots;
• Suggestion; the suggestion contains the following information:
– relative number of TRX;
– bit map containing 1 for each timeslot selected by the PCU in the suggested
configuration (there is 1 for both pre-allocated and idle channels in the configura-
tion);
– bit map containing 1 for each timeslot pre-allocated by the PCU in the suggested
configuration;
Pre-allocated channels are the channels already in packet transfer
i mode, but without assigned TBF; i.e., they are the channels for which the
TEMPCH timer is still running (see "5.3.3.1 PCU Algorithm").

• The HA/VA indicator. This indicator is used to indicate in which allocation type
(HA/VA) the PCU has sent the message to the TDPC;
• Number of needed Abis subslots for each PDCH.
As a general rule, the TDPC will first try to satisfy the suggestion sent by the PCU. Only
if it is not possible to exactly satisfy the suggestion, it tries to satisfy the request using
as many pre-allocated channels as it can. If again the request is not satisfied, the TDPC
goes on to search through all the TRXs, in order to find out the best configuration that
matches the requirement fixed further.
It is important to underline the following feature: Abis/PDT scarcity does not affect the
radio resource assignment algorithm of TDPC. The only mandatory check (on TDPC)

100 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

concerns the availability of one Abis/PDT per new PDCHs in the selected radio timeslot
configuration. No attempt is done to search radio resources minimizing the number of
new allocated Abis/PDT resources. Hence, in case of Abis/PDT resources scarcity it is
not guaranteed that the initial coding scheme can be supported; and the initial target
throughput is based on the number of radio timeslots that can be actually activated.
Then the TDPC will answer to the PCU with:
– a PDCH_Setup message when at least one idle channel has been assigned; in this
case, no matter of the value of the thresholdIdleAbisStopUpgrade field of the
GASTRABISTH parameter (see "5.3.4.2 Upgrade of Abis Resources"), the TDPC
will allocate new PDCHs trying to assign them the requested number of Abis/PDTs
per PDCH and, if necessary and possible (see "5.3.4.2 Upgrade of Abis
Resources"), upgrade up to the requested number of PDTs per PDCH the already
allocated PDCHs in the configuration.
When Abis/PDT resources are not enough to completely satisfy the request (activa-
tion of new PDCHs and possible upgrade of already allocated PDCHs), the number
of PDTs per PDCH specified in the request is downgraded.
– a PDCH_KO message, if no idle channels have been assigned (even if some pre-
allocated channels were present in the PCU request); also in this case, if necessary
and possible (see "5.3.4.2 Upgrade of Abis Resources"), upgrade up to the
requested number of PDTs per PDCH the already allocated PDCHs in the configu-
ration.
The TDPC algorithm is described in Fig. 5.8.
When the involved BTS is congested, if the incoming GPRS/EGPRS request is for more
than one timeslot, the TDPC distinguishes between upgrade requests and new
requests:
– an upgrade request is detected each time the PCU requires additional timeslots for
GPRS/EGPRS service. This means that some timeslots are currently allocated for
PS data transmission, and the request is for additional resources. In this case, when
the BTS is congested, the request from the PCU is rejected and the TDPC sends a
PDCH_KO message to the PCU.
– a different situation occurs when an incoming request arrives at the TDPC from the
PCU and no channels are currently allocated for PS services. In this case, when the
BTS is congested, the incoming multislot request is downgraded to a single timeslot
request. At this time if the request cannot be served immediately, it will be included
in the waiting queue.
This mechanism is not applied to timeslots reserved for exclusive use of the
i GPRS/EGPRS services. So if the incoming request can be satisfied using the timeslots
reserved exclusively for PS services (fixed by the operator using the GMANPRES
attribute) no downgrade or reject is performed on the incoming request.

If the BTS is not congested, the TDPC verifies if there are some pending requests, first
in the queuing list and then in the waiting queue. If any pending request exists then the
TDPC puts the incoming GPRS/EGPRS request in the waiting queue because it is
necessary to serve the old calls first.
So, when resources are available and either the queuing list or the waiting queue is filled
with some pending request, the new request will not be served immediately, even if
there is no congestion from a BTS point of view. This is done in order to optimize the
usage of resources and it can produce a short delay in serving the new requests.

A30808-X3247-L24-5-7618 101
GPRS/EGPRS Global Description Information
System

In case both the queues are empty, the TDPC has to check if the incoming request can
be completely satisfied by available system resources.
The algorithm on TDPC will search idle channels following these criteria:
1. prefer EGPRS on EDGE TRXs and GPRS on non-EDGE TRXs;
2. maximizing the initial target throughput;
3. using as many pre allocated channels, if any, as it can (resulting from PCU sugges-
tions);
4. minimizing the number of forced intracell handovers of circuit switched calls;
5. choosing the preferred TRXs according to the CPOLICY parameter.
If the request is completely satisfied by the available resources, without the need to
execute forced intracell handovers, the request is served immediately; so the TDPC will
answer to the PCUC with a PDCH_Setup message. The PDCH_Setup message always
contains the current allocation value (VA/HA) on TDPC.
If one or more intracell handovers have to be executed, the request is put in waiting
queue and the management is delegated to the waiting queue manager process (see
"5.3.6 Waiting Queue Management").
Note that for the previous algorithm, the search including forced intracell handovers is
i applied only if forced intracell handovers have been enabled by the operator (see
"5.3.6.3 Forced Intracell Handovers of Already Established CS Calls").

If no new idle channels are assigned, the TDPC will answer to the PCU with a
PDCH_KO message; this message has a field as a bit map containing the HA/VA indi-
cator.
The HA/VA indicator is set to horizontal allocation or vertical allocation depending on the
situation of radio interface and Abis interface, described in "5.3.2 Horizontal/Vertical
Allocation Strategies".
The following considerations can also be done:
• each time more than one solution is found to satisfy a request, it is chosen that for
which, when new channels are assigned, the number of adjacent busy channels for
GPRS/EGPRS is higher. This is done to reduce holes in the configuration and to
facilitate the assignment for new incoming GPRS/EGPRS calls when the VA is
active;
• it should be noted as the priority related to the preferred TRX is the lowest one; so
if the request can be satisfied, according to the other criteria, on not preferred TRXs,
the resources will be assigned on a not preferred TRX;
• in case more than one allocation with the same number of timeslots is possible on
different TRXs, the allocation is performed according to the order of priority listed
above.
For instance if the system is handling a request for three timeslots, and both TRX0
(BCCH) and TRX1 (non BCCH) have three available timeslots, but only TRX0 has
one “empty channel”, whereas TRX1 has no empty channels, then the allocation is
performed on TRX0 even if TRX1 may have more than the required timeslots free;
• for PDCH allocation in multislot configurations, the allocated PDCHs must have (see
also "4.7 Multislot Configuration"):
– same frequency hopping law;
– same training sequence code (TSC);
– same MAIO;

102 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

– adjacent time slot numbers.


The first three rules have to be followed during the configuration phase, that means
that all timeslots (defined as TCH) must have the same hopping law and the same
training sequence code.
The fourth law is dynamically followed by TDPC, when it selects timeslots to be allo-
cated to GPRS/EGPRS users.
• if the TDPC answers with PDCH_KO to the PCU because no GPRS/EGPRS chan-
nels are available or because no idle channels are assigned and there are no pre
allocated ones, the TDPC will force the VA. In this way the PCU will try to layer the
call on already busy for GPRS/EGPRS channels;
• the subsequent deallocation of the allocated PDCH occurs when all the TBFs
present on it have been released, and the TEMPCH timer has expired.
If what is assigned by the TDPC does not fit what is required by the PCU, this last tries
to expand the proposed configuration using timeslots already available and adjacent to
the new ones. For instance when the HA is used, if the PCU has to assign to a
GPRS/EGPRS user three channels, the PCU requires three idle channels to the TDPC
(let us suppose that the PCU does not indicate any pre allocated channel in the sugges-
tion).
If the TDPC can assign only two channels, because either the maximum number of PS
channels or the Vertical allocation threshold has been reached, it assigns these two
channels and also sets the VA/HA flag to VA.
The PCU then uses the two channels assigned by TDPC, plus another channel already
available at PCU side where another TBF is running (since there are not any pre allo-
cated channels).

A30808-X3247-L24-5-7618 103
GPRS/EGPRS Global Description Information
System

Reception of
PDCH_Request
from PCU

YES Is the Cell NO YES


Are queues
congested?
empty?

NO

Put the call in


Waiting Queue
Upgrade New call
New call or Are there
NO available YES
Upgrade? resources
immediately?

Assign the resources


Downgrade the (E)GPRS
call to 1 timeslot Is it
NO possible to free
resources using
Intracell Handover? Calculate HA/VA
Send PDCH_KO threshold
Put the downgraded
to PCU
call in Waiting Queue YES
(if necessary set VA)

Put the multislot YES NO


Threshold
call in Waiting Queue overcome?

Were there
Set VA
YES pre-allocated channels NO
in the PCU request?

Assign pre-allocated Calculate HA/VA


PDCHs threshold
Send PDCH_KO Send PDCH_Setup
to PCU to PCU
(Set VA)
YES Threshold NO
overcome?

Set VA

Send PDCH_KO
to PCU

Fig. 5.8 Allocation Algorithm followed by the TDPC

104 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

5.3.4 Upgrading Strategies


Two kinds of upgrading strategies are defined according to different situations:
1. upgrading of radio resources: additional radio resources could become necessary
to sustain the peak throughput;
2. upgrading of Abis/PDT resources: it can be required by Link Adaptation (see 10.5)
when conditions get better, or by the PCU when in the same PDCH has to fill a TBF
with an higher coding scheme with respect to those already substained.
As it can be seen, the upgrading of radio resources is different from the upgrading of
Abis/PDTs, that occurs under completely different conditions. So, these upgrading strat-
egies are discussed separately.

5.3.4.1 Upgrade of Radio Resources


After the first allocation of radio resources, additional radio resources could become
necessary to sustain the peak throughput. Therefore a radio resource upgrading
strategy is necessary. The events triggering the radio resource upgrading (upgrading
conditions) are:
– increasing of the peak throughput requirement;
– decreasing of the “maximum sustainable throughput”, due to the worsening of radio
conditions;
– loss of radio resources due to pre-emption or O&M commands.
In general, once detected the upgrading condition, several additional radio resources
could be necessary to fill the gap to the required number of radio resources. However,
in the present release, for effort reasons, the upgrading of radio resources is performed
one step a time; it means that radio resources are added one at a time, each time one
of the upgrading conditions is detected. The additional radio resource must belong to
the same TRX and must be adjacent to the radio resources already assigned to the TBF.
The choice between the radio resource on the left or on the right of the current allocation
is performed using the same criteria used in the first allocation of radio resources
(see 5.3.3.1).
Note that in case the worsening of radio conditions would lead simultaneously to a step
of Link Adaptation (downgrading the CS/MCS and possibly the Abis/PDTs, see
"10.5 Link Adaptation") and to upgrading of radio resources, the downgrading of
CS/MCS is managed before the upgrading of radio resources. It is a general rule on
PCU that procedures cannot be nested: hence the upgrading of radio resources will be
started, if necessary, only when the downgrading of CS/MCS procedure is completed.
The radio resources upgrade is attempted if the already allocated resources are less
than what can be supported by the MS multislot class.
Depending on the position of the TBF on the TRX, the best additional PDCH could be
already allocated to GPRS/EGPRS, or it could be necessary to request it to TDPC. As
long as vertical allocation is in progress, the PCU is not allowed to request new PDCHs
to TDPC for upgrading reasons.
The additional PDCH can be requested only if both the following conditions are satisfied:
1. the horizontal allocation is in progress;

A30808-X3247-L24-5-7618 105
GPRS/EGPRS Global Description Information
System

2. the number of idle radio timeslots in the cell is higher than the thresholdIdleChanEn-
ableUpgrade field of the GASTRTH attribute.
The number of idle timeslot is calculated in the same manner as described in
i "5.3.2.3 Switching between VA and HA According to Radio Conditions".

The check is performed on the TDPC. The PCU is informed by a flag (enableRadi-
oUpgradingFlag) added in all the messages containing the allocation status flag. At
system initialization, by default, the enableRadioUpgradingFlag is DISABLED both on
PCU and TDPC sides, and is set to ENABLED at the first check detecting the horizontal
allocation condition, unless the thresholdIdleChanEnableUpgrade value is 100 (this
value means: new PDCHs cannot be allocated to GPRS for upgrading reasons).
The thresholdIdleChanEnableUpgrade does not enable the ‘upgrading strategy’. It
enables the possibility to allocate new PDCHs to GPRS/EGPRS for upgrading reasons.
But PDCHs already allocated to GPRS/EGPRS can be assigned to a TBF for upgrading
reasons no matter of the thresholdIdleChanEnableUpgrade value. Besides, the thresh-
oldIdleChanEnableUpgrade threshold does not affect the assignment of resources for
new incoming TBFs.
In the following the upgrading conditions are discussed.

1) Change in Peak Throughput Requirement


While an uplink TBF is in progress, the mobile can request a variation in peak throughput
submitting a PACKET_RESOURCE_REQUEST message and specifying a different
value of peak throughput in the ‘Channel Request Description’ information element.
While a downlink TBF is in progress, the SGSN can request a variation in peak
throughput specifying a different value of throughput in the ‘QoS Profile’ information
element of the DL-UNITDATA.
In both cases (DL/UL), the variation of peak throughput is taken into account only if a
peak throughput value higher than the currently managed value (in the same UL or DL
direction) is specified.
An extension to the number of allocated timeslots is tried if the number of currently allo-
cated timeslots is lower than the number of required timeslots; the number of required
timeslots is defined as:

Number of required TSs = min (ceil ( new PT / (T_A_CS x (1-BLER)), Multislot Class).

where:
ceil = round up to the upper integer
new PT = new required Peak Throughput
T_A_CS =throughput of the Actual Coding Scheme
BLER = it is the actual BLER.

The extension is tried by adding one adjacent TS to the actual configuration; so the PCU
will send to TPDC a PDCH_Upgrade_Reqeust message, but only if the conditions
regarding horizontal allocation and the percentage of idle timeslots are verified.
In case radio resources are missing and the upgrade is not possible, the upgrading
request is dropped. The upgrading will be attempted again if a decreasing of maximum
sustainable throughput is detected, as specified in 2) Change in “Maximum Sustainable
Throughput”.

106 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

2) Change in “Maximum Sustainable Throughput”


During a TBF lifetime, due to variations in radio conditions, either the BLER or the used
CS/MCS coding scheme can change, leading to a change in ‘Maximum sustainable
Throughput’.
The Maximum sustainable throughput is defined as the maximum throughput that would
be achieved by a given TBF if it was alone on the multislot configuration, that is:

Maximum sustainable throughput = T_A_CS x (1-BLER) x #TS

where:
T_A_CS =throughput of the Actual Coding Scheme
BLER = it is the actual BLER
#TS = number of allocated timeslots to the TBF

A check on the maximum sustainable throughput is performed periodically, with a period


defined by the UPGRFREQ attribute.
As a general rule, only the decreasing in maximum sustainable throughput is taken into
account (increasing, hence the ‘downgrading’ of radio resources is not managed). More-
over, since the variations in Maximum sustainable throughput can be very frequent, only
the decreasing below a given threshold will be managed.
An extension to the number of allocated TSs is tried if:

T_A_CS x (1-BLER) x Currently allocated #TS < (1- ACCEPTGDEGR) x PT

where:
T_A_CS = throughput of the Actual Coding Scheme
BLER = it is the actual BLER
PT = Peak Throughput
ACCEPTGDEGR= it is an O&M parameter
So, when the maximum sustainable throughput becomes lower than the maximum toler-
able degradation of the peak throughput, the upgrade is performed.
As long as the ‘one radio resource a time’ algorithm is implemented, the ACCEPT-
i GDEGR attribute is suggested to be set to 0 (no degradation allowed, radio resource
upgrading always attempted as soon as the upgrading condition is detected), in order
to reach the required radio resource allocation in several steps.

The extension is tried by adding one adjacent timeslot to the actual configuration; so the
PCU will send to TPDC a PDCH_Upgrade_Reqeust message, but only if the conditions
regarding horizontal allocation and the percentage of idle timeslots are verified.

5.3.4.2 Upgrade of Abis Resources


According to flexible Abis allocation strategy, for already assigned radio channels, new
Abis resources could become necessary, e.g.:
– it can be required by Link Adaptation when the radio conditions gets better (see
"10.5 Link Adaptation");
– it can be required by the PCU when in the same PDCH has to set up a TBF with a
higher coding scheme with respect to those already multiplexed on the PDCH (see
"5.3.3 Management of Incoming GPRS/EGPRS Requests").

A30808-X3247-L24-5-7618 107
GPRS/EGPRS Global Description Information
System

The user can manage the upgrade strategy of Abis resources by two fields of the
GASTRABISTH parameter. The two fields are:
• thresholdIdleAbisStopUpgrade field: it defines the percentage of idle Abis subslots
of a BTSM (over the available Abis subslots managed by the BTSM) under which
the PCU must disable the Abis upgrading requests to TDPC for all the cells
managed by the PCU and belonging to the involved BTSM. When this threshold is
overcome, the first allocation of Abis resources to a TBF is performed with the same
criteria used under normal conditions (looking at the candidate initial coding
scheme), but further upgrading of Abis resources is forbidden. Moreover, in case of
runtime Abis release (due to worsening of radio conditions, CS pre-emption or O&M
commands), the released Abis is not allowed to be allocated again to running TBFs.
The main aim of this threshold is to avoid useless signalling between PCU and
TDPC in case of nearly complete Abis congestion, therefore, the default value of the
threshold is 0, meaning that the Abis upgrading is disabled only in case of complete
Abis congestion. The secondary aim of this threshold is to avoid the allocation of
additional Abis resources to running packet services in case of Abis scarcity, so that
the residual Abis resources in the pool can be by preference available to set up new
CS services (this will be the trend in case of vertical allocation) or even to new PS
services (in case horizontal allocation is still active). Note that moving this threshold
from the default value, a reduction in PS throughput is expected;
• thresholdIdleAbisRestoreUpgrade field: it defines the percentage of idle Abis
subslots of a BTSM (over the available Abis subslots managed by the BTSM) over
which the Abis upgrade requests to TDPC are restored for all the cells managed by
the PCU and belonging to the involved BTSM.
Constraints on the Abis thresholds are:

thresholdIdleAbisStopUpgrade < thresholdIdleAbisRestoreUpgrade

There is no constraint between the Abis threshold to switch to vertical allocation (see
i "5.3.2.4 Switching between VA and HA according to Abis Interface Conditions") and the
Abis threshold to disable the ‘Abis upgrade requests’; the operator is free to set the one
lowest than the other, and vice versa.

As a general configuration rule, in BTSMs where the Abis resources/radio resources


ratio is quite high, in order to obtain the highest benefit from Link Adaptation, the Abis
upgrading should be disabled only in case of complete Abis congestion or at least after
the switch to vertical allocation.
Instead, in BTSMs where the Abis resources/radio resources ratio is quite low, for some
operators it could be preferable to disable the Abis upgrading before the switch to
vertical allocation.
In any case, the choice will depend on the relative preference given from the operator
to circuit switched calls versus packet switched TBFs, and to running TBFs versus
incoming TBFs.

5.3.5 Incoming CS Calls


When a circuit switched call (deriving from either a normal assignment or an external
incoming handover) comes into the cell with no free radio channels, the following algo-
rithm is applied:

108 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

1. if the incoming CS call finds the cell in a congested state, the first attempted task is
to preempt one vulnerable CS call;
2. if preemption cannot be started for whatever reason (e.g., the feature is not
enabled), the Directed Retry procedure is started;
3. if also the Directed Retry cannot be started (e.g., the feature is not enabled or the
feature is enabled but the Handover Condition Indication message does not contain
any cell) the queueing procedure is started, if enabled;
The queuing procedure puts the CS call in the Queueing List that is
i different from the Waiting Queue.

4. if the queueing procedure is not enabled, the CS call is put in the Waiting Queue.
To free resources for the CS call put in waiting queue, a packet data transfer may be
downgraded, in cases of a multislot call, or released, in the worst case. In any case,
the static GPRS/EGPRS channels can not be pre-empted by CS calls (see
"5.3.6 Waiting Queue Management").
According to the flexible Abis allocation strategy (see 6.3), it could also happen that
when the CS calls have to be served, no Abis resources are available to serve the
incoming call. Even in this case, the call is put in the waiting queue in order to find the
required resources.

5.3.6 Waiting Queue Management


As it has been described in "5.3.3.2 TDPC Algorithm", there are some cases for which
the TDPC inserts the incoming requests in the waiting queue; then the TDPC must
analyze a second time these pending requests to serve them. The task is activated by
a timer.
To summarize, the following GPRS/EGPRS calls are put in waiting queue:
a) GPRS/EGPRS requests that arrive when the cell is congested and no GPRS calls
are present in the cell; these calls are downgraded to 1 timeslot before being
inserted in the waiting queue;
b) GPRS requests that arrive when other calls are present in the waiting queue or in
the queuing list;
c) GPRS requests that must wait for intracell handovers; in these cases, the multislot
call is inserted in the waiting queue.
CS calls that arrive:
a) when no radio resources are available in the BTS; in fact, if both the pre-emption
and the directed retry fail or have not been enabled, these calls are put in the
queuing list (if the Queuing feature has been enabled).
Otherwise, if the Queuing feature is not enabled, the CS calls are put in the waiting
queue (see "5.3.5 Incoming CS Calls")
b) when no Abis resources are available in the BTSM that manages the BTS.
could also be put in the waiting queue.
The TDPC, before checking the waiting queue, analyses the queueing list. If the
queueing list contains some pending request (i.e., some CS calls), the TDPC will imme-
diately manage resources of the queueing list.

A30808-X3247-L24-5-7618 109
GPRS/EGPRS Global Description Information
System

After this procedure, or if the queueing list is empty, the process will analyzes the waiting
queue.
Note, once more, that the resources that are released, are used first by the Queueing
i process and only later on by the Waiting Queue process. So the classic Queueing
procedure already implemented always has a higher priority than the waiting queue
management.

Three types of action can be performed by the process to serve pending requests on the
waiting queue:
1. Use resources just released by the TDPC: in case the system had released any
system resources, these have been included in the Idle List structure. Then the
system finds the released resources which are available for the specific cell. If the
resources are not enough to serve all the entries present in the waiting queue, the
following Downgrading mechanisms are activated;
2. Downgrading of already active HSCSD multislot calls: the downgrade of already
active HSCSD calls, is performed in two situations only:
– to serve GPRS/EGPRS pending requests in the waiting queue;
– to serve incoming CS requests in the waiting queue (see "5.3.5 Incoming CS
Calls");
The downgrade of an already active HSCSD call is executed only if the number of
used timeslots is greater than one (i.e., at least one timeslot must remain allocated
for the HSCSD call)
3. Downgrading of already active PS multislot calls: the GPRS/EGPRS downgrade
process, consists in a decrease of the number of timeslots already assigned to PS
services. When the downgrade of PS calls is performed, one of the GPRS/EGPRS
channels is “preempted” and the channel is released. In the case in which a PS data
transmission uses only one timeslot for GPRS/EGPRS, and the timeslot is
preempted for downgrading, the transmission is interrupted (to avoid GPRS/EGPRS
downgrading, the operator can assign static GPRS/EGPRS timeslots as explained
in "5.2 Configuration of GPRS Channels in a Cell"). As it has been described (see
"5.3.5 Incoming CS Calls"), the downgrade of already active PS multislot calls is
performed to serve incoming CS requests in the waiting queue.
No active GPRS/EGPRS calls are downgraded to free resources for incoming
i GPRS/EGPRS calls.

Regarding the downgrade of already active GPRS/EGPRS and HSCSD multislot calls,
the user can select the downgrade strategy. The user can choose the preferred down-
grade strategy through the DownGradeStrategy parameter (DGRSTRGY). This
attribute allows the user to choose among five different strategies:
– Downgrade of HSCSD calls first
– Downgrade of GPRS/EGPRS calls first
– Downgrade of HSCSD calls only
– Downgrade of GPRS/EGPRS calls only
– No Downgrade

5.3.6.1 Pre-emption of PDCH Channels


The preemtpion of a PDCH is executed to serve an incoming CS call (see
"5.3.5 Incoming CS Calls"); the TDPC can send to the PCU the order to release one or

110 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

more PDCHs in a cell (PDCH_preemption_order), due to unavailability of radio


resources in a cell.
On the PCU, PDCH pre-emption orders are sent:
1. first to PDCHs with the ‘PDCH empty timer’ still active; as described, the “PDCH
empty timer” can be defined by the user via the TEMPCH parameter;
2. then to PS dynamic channels.note that:
Note that:
– before downgrading a PDCH on a TRX where CS calls are seen as preferred, a free
channel on not preferred TRXs (not preferred for CS calls) is searched first (i.e., the
CPOLICY parameter is negletted);
– independently on the value of the DGRSTRGY parameter, if there are some empty
channels (i.e., PDCH channels for which the PDCH empty timer is running), they are
used to serve the CS call; if no empty channels exists, then the rules defined by the
DGRSTRGY parameter are followed to free resources for the CS call. The down-
grade of the PDCH channels is applied, until the number of the reserved channels
for PS services is reached in the cell (reserved PS channels cannot be down-
graded).
The preemption of some PDCHs can result in the reconfiguration of some TBFs (also
partially allocated on ‘residual’ PDCHs) and in the release of some other TBFs
(completely allocated on the preempted PDCHs).

5.3.6.2 Pre-emption of PDT Resources


The TDPC can send to the PCU a PDT pre-emption order in case no PDT or Abis
resources are available for PBCCH/PCCCH activation, or in case no Abis resources are
available to serve incoming CS requests put in the waiting queue.
Cells belonging to the ‘exhausted’ BTSM are spread over several PCUs (see "8 Load
Control for Packet Switched Services"), so, in case of unavailability of Abis resources in
a BTSM pool, the TDPC selects the PCU with the highest number of PDTs allocated to
the ‘exhausted’ BTSM, and sends it a pre-emption order.
In case of unavailability of Abis resources, the same ‘pre-emption level’ used when radio
resources unavailability is applied. For example, in case of CS pre-emption due to
unavailability of Abis resources, the pre-emption management will not cause the release
of PBCCHs and PCCCHs (or the pre-emption of static GPRS/EGPRS channels). That
is: only PDTs with SFC=1..4 can be removed from PBCCHs and PCCHs.
On PCU, PDT pre-emption orders are managed with the aim to balance the distribution
of PDTs among cells and to disturb as little as possible the running TBFs. In each
selected cell, the algorithm selects, in the order:
1. PDTs with ‘PDT empty timer’ still active (as it has been described in
"5.3.1 Generalities about Resource Assignments", the “PDT empty timer” can be
defined by the user via the TEMPPDTparameter);
2. starting from the first TRX and highest timeslot numer and removing one PDT per
timeslot, then moving on the next TRX; this step is repeated until the required
number of PDTs is found.
Since the PDT pre-emption management can result in the release of some PDCHs, it
can result in the reconfiguration of some TBFs (partially allocated also on ‘residual’
PDCHs) and in the release of some other TBFs (completely allocated on the preempted

A30808-X3247-L24-5-7618 111
GPRS/EGPRS Global Description Information
System

PDCHs). PDCHs can be released provided that the number of residual PDCHs allo-
cated in the cell is higher or equal to the number of reserved PDCHs.
In case the release of the PDTs does not cause the whole PDCH release, a forced
downgrade of the coding scheme is performed for all the TBFs multiplexed on the
involved PDTs. Otherwise (release of the whole PDCH), the same behavior imple-
mented in case of PDCH pre-emption (see 5.3.6.1) is ensured.

5.3.6.3 Forced Intracell Handovers of Already Established CS Calls


When a request comes from the PCU for a new PDCH, the TDPC tries to allocate it.
It could happen that in the transceivers supporting GPRS/EGPRS there are not free and
consecutive timeslots, e.g., because the already assigned PDCHs are full and the
remaining timeslots are dedicated to circuit switched calls. In this case, if there is at least
one free channel in the cell (and the maximum number of timeslots assigned to PS
services has not been reached), a forced intracell handover starts to free timeslots for
GPRS users. The forced intracell handover allows the moving of a CS call from one
timeslot, to another one in the same cell.
The decision whether pre-emption may be made on circuit switched services is taken
i by the TDPC, causing a forced intracell handover for circuit switched calls.
To enable forced intracell handover, the user must set the ENFOIAHO parameter
(Enable Forced Intracell Handover) to TRUE.

112 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

6 Hardware and Software Architecture


For the management of Packet Switched Services in the SBS system, the Packet
Control Unit (PCU) has been designed for the BSC. It supports the packet data inter-
working between the “Gb” interface and the “Abis” interface.
The functionalities of the PCU are supported by the Channel Codec Units (CCUs) which
have been implemented in the BTSs (see the "Fig. 6.1 Hardware and Software Entities
supporting the GPRS/EGPRS technology"). The CCU software is downloaded to the
BTS from the BSC.

Fig. 6.1 Hardware and Software Entities supporting the GPRS/EGPRS technology

The PCU unit within the BSC provides the following functions:
– Channel Access Control functions, for example access requests and grants;
– PDCH scheduling functions for uplink and downlink data transfer;
– Radio Channel Management functions, like power control, congestion control,
broadcast control information, etc.;
– PDCH RLC ARQ functions, including buffering and re-transmission of RLC blocks;
– LLC layer PDU segmentation into RLC blocks for downlink transmission;
– RLC layer PDU re-assembly into LLC blocks for uplink transmission;
– management of the protocols supporting the “Gb” interface.
The CCU unit within the BTS provides the following functions:
– Channel coding functions, including FEC and interleaving;
– Radio channel measurement functions, including received quality level, received
signal level and information related to timing advance;
– Continuous Timing Advance update.
The PCU functional managed object models the physical packet control unit designed
to implement the packet switched services (PS) in the SBS system.

Functional object Meaning

PCU This Functional Managed Object (FMO)


represents the Packet Control Unit designed
to implement GPRS services in the SBS
system. Depending on the BSC type this
FMO can be implemented by means of the
boards PPCU (for the standard BSC) or
PPXU (for the High Capacity BSC).

Tab. 6.1 PCU Functional Managed Object (FMO)

A30808-X3247-L24-5-7618 113
GPRS/EGPRS Global Description Information
System

In case of PPCU boards for the standard BSC each PCU can handle up to 64 PDTs on
the Abis interface with a maximum number of 16 FRLs on the Gb interface. FRL is the
Functional Managed Object that models the physical link connection on the “Gb” inter-
face. The connection can be realised through the A interface (PCMA link) or directly to
the SGSN through the PCMG link. The Packet Data Terminal (PDT) represents a basic
16 kbit/s resource for the packet switched services manageable by the PCU.
In case of PPXU boards for the High Capacity BSC each PCU can handle 256 PDTs on
the Abis interface and up to 64 FRLs on the Gb interface.
In the next chapters it will be described in detail the difference that the support of the
GPRS/EGPRS technology requires in terms of hardware supported and software appli-
cations to the standard and High Capacity BSC and also to the different BTS types.

6.1 Supported BSC Types


In the current BR 7.0 release the BSCs that can be installed in the Siemens customers’
mobile networks are the following:
1. the BSC equipped with the SN16 switching matrix and with older peripheral proces-
sors called “standard BSC” (see the chapter: "6.1.1 “Standard” BSC").
2. the High capacity BSC based on the same rack as the “standard” one, but equipped
with the new SNAP switching matrix called “High capacity BSC with the old rack”
(see the chapter: "6.1.2 High Capacity BSC with the Old Rack"), that provides better
performances in terms of:
– connectivity (i.e., number of supported PCM lines).
– packet data handling capability.
– LAPD signalling.
3. the High capacity BSC based on a new rack called “High capacity BSC with the new
rack” (see the chapter: "6.1.3 High Capacity BSC with the New Rack"), that provides
better capabilities with respect to the “High capacity BSC with the old rack” due to
the installation of the new LICD cards.
According to the maximum number of supported PCM line interfaces, High Capacity
i BSC with Standard Rack is also called HC BSC 72, whereas High Capacity BSC with
new Rack is called HC BSC 120.

114 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

In the next paragraphs the different BSC types are described, taking into particular
i account their hardware and software resources configured for supporting the GPRS and
EGPRS technology. It is also important to make a distinction between the terms “Packet
Data Channel (PDCH)” and “Packet Data Terminal (PDT)”.
The Packet Data Channel (PDCH), as it has been described in the chapter "4 Radio
Interface Description", is the radio timeslot associated to packet switched services (that
means when the timeslot is associated to the packet switched services, it is called
PDCH).
The Packet Data Terminal (PDT) represents a basic 16 kbit/s resource for the packet
switched services manageable by the PCU. The capacity of the PCU, from packet
switched data services point of view, is assigned in terms of Packet Data Terminals,that
means that a PCU supports a certain number of Packet Data Terminals. This number
of Packet Data Terminals corresponds to the number of Abis subslots (16 kbit/s)
manageable by the PCU.For example, when a single PDCH is associated to a GPRS
user using CS1 coding scheme, it is also associated to a single Abis subslot, and so only
one PDT is busy in the PCU that manages this PDCH (in this case, there is a one to one
relationship between PDCH and PDT); but when a single PDCH is associated to an
EGPRS user using MCS9 coding scheme, five Abis subslots are associated to this
PDCH (see the chapter: "6.3 PCU Frames and Dynamic Allocation on the Abis Inter-
face") , and so five PDTs are busy in the PCU that manages this PDCH (in this case,
there is a one to five relationship between PDCH and PDT).

6.1.1 “Standard” BSC


In the Standard BSC, the PPCU processors are used to implement the PCU unit; each
PCU unit consists of two PPCU boards:
– one of them is in Providing Service state;
– the other one is in Cold Stand-by state, and it is used as a spare board.
The PPCU boards are inserted in the BSC rack in place of some PPLDs, as it is shown
in the Fig. 6.2.

A30808-X3247-L24-5-7618 115
GPRS/EGPRS Global Description Information
System

Fig. 6.2 View of the BSC Rack with and without PPCU Boards.
Two instances of the PCU object can be created:
– PCU:0;
– PCU:1.
The creation of one PCU object implies the consequent creation of the two related
PPCU boards (active and cold-standby):
• the creation of the PCU:0 involves the creation of both the PPCU:0 and the PPCU:1;
• the creation of the PCU:1 involves the creation of both the PPCU:2 and the PPCU:3.
The system firstly creates the two PPCU objects, and then the PCU object. After the first
card reaches the Providing Service state, the PCU starts the configuration alignment.
Since the PPCUs are inserted in the BSC rack in substitution of some PPLDs, when the
user creates a PCU object instance, some PPLDs may not be equipped. The rule is
shown in Tab. 6.2.

116 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

PCU instance PPLDs to be


removed

PCU:0 PPLD:11
PPLD:12
PPLD:13
PPLD:14
PCU:1 PPLD:7
PPLD:8
PPLD:9
PPLD:10

Tab. 6.2 PPLD Boards to Be Removed according to the PCU Object Instance.

Each PPLD board can manage up to 8 LAPD channels; so, when 14 PPLD boards are
used, 112 LAPD channels are available in the BSS. When some PPLD boards are
removed to introduce PPCU boards, the number of LAPD channels decreases, reducing
the signalling capability of the BSC.
When only one PCU is created, the number of PPLD boards becomes 10, and the
number of configurable LAPD channels is 80; when both the PCU instances are created,
the number of PPLD boards decreases to 6, and the number of configurable LAPD
channels is 48.
Each Packet Control Unit is able to handle at most a data rate of 2 Mbit/s. This data flow
is divided in two data rates of 1 Mbit/s each one:
1. a data rate of 1 Mbit/s towards the Abis interface; this flow allows the management
of the Abis interface at most 64 GPRS channels (16 kbit/s each one), i.e., 64 PDTs;
2. a data rate of 1 Mbit/s towards the Gb interface; this flow allows the management of
the Gb interface at most 16 Frame Relay Links (64 kbit/s each one, see "7 Gb Inter-
face").
When the standard BSC is fully equipped with two PCUs, it can handle up to 128 GPRS
channels (PDCHs).
EGPRS is not supported by the standard BSC due to its low capacity in terms of PDTs.
i
For each BSC, it is possible to configure up to 150 cells and, as a consequence, up to
150 PTPPKF object instances.

A30808-X3247-L24-5-7618 117
GPRS/EGPRS Global Description Information
System

6.1.2 High Capacity BSC with the Old Rack


Using the same rack of the previous releases, it is possible to get a BSC with a higher
capacity by changing some boards (see Fig. 6.3).

Fig. 6.3 View of the “High Capacity” BSC with the Traditional Rack.

In order to get a High Capacity BSC using the traditional rack, the following boards are
used:
• a new switching matrix board, called SNAP;
• new peripheral processor boards i.e.;
– PPXL boards to manage both LAPD and SS7L signalling;
– PPXU boards to manage GPRS and EGPRS services.
It is important to underline that the hardware of both PPXLs and PPXUs is named
i PPXX; depending on the slot position inside the BSC rack, the same board (i.e., the
PPXX one) acts, from the functionality point of view, as PPXU or PPXL.

In comparison with the SN16 (i.e., the switching matrix of the “standard” BSC), the
SNAP card allows the interface of 48 lines at 8 Mbit/s coming from LICD and PPXX
(double bandwidth in comparison with the SN16, which can interface 24 lines).

118 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

This doubled number of lines increases independently (i.e., without trade-off) both
GPRS/EGPRS and LAPD channels.
The new switching matrix is introduced in the system through the handling of the
NTWCARD attribute; this attribute can assume the values:
– NTWSN16, when SN16 switching matrix is used (standard BSC);
– NTWSNAP, when the SNAP switching matrix is used (high capacity BSC with the
old rack).
When the NTWCARD is set to NTWSN16, the BSC works with PPCC, PPLD and PPCU
i boards.
When the attribute value is NTWSNAP, only the SNAP and the new PPXU and PPXL
boards are allowed.
Mixed configurations are not possible.

In the high capacity BSC, to get more GPRS/EGPRS channels it has been necessary
to increase the number of boards assigned to packet switched functionality and to
increase also their capability. This is allowed by the SNAP switching matrix, which
provides 8 lines at 8 Mbit/s towards PPXXs. Two lines are used for handling LAPD and
SS7 level 2 signalling protocols with the new PPXL boards; and the remaining six are
used for PPXU boards (each PPXU board has its own 8 Mbit/s line).
The PPXUs are all placed in the extended rack (as it is for the PPCUs).
To handle packet switched services, six instances of the PCU object can be created:
– PCU:0;
– PCU:1;
– PCU:2;
– PCU:3;
– PCU:4;
– PCU:5.
The creation of one PCU object implies the consequent creation of one PPXU board: i.e
the creation of the PCU:0 involves the creation of the PPXU:0; the creation of the PCU:1
involves the creation of the PPXU:1, and so on.
A PPXU card is automatically created when the PCU object with the same instance is
i created, if NTWCARD= NTWSNAP. Otherwise (if NTWCARD=NTWSN16), a couple of
PPCUs are created (PPCU 0,1 for PCU-0; PPCU 2,3 for PCU-1).

Tab. 6.3 shows the correspondence between the boards of the “standard” BSC and
those of the high capacity BSC from packet switched services point of view.

Standard BSC (no Standard BSC (full High capacity


GPRS) GPRS configuration) BSC

PPLD-3 PPLD-3
PPLD-4 PPLD-4 PPXU-0
PPLD-5 PPLD-5
PPLD-6 PPLD-6 PPXU-1
PPLD-7
PPLD-8 PPCU-2 PPXU-2

Tab. 6.3 Correspondence between the Boards of the Two Types of BSC

A30808-X3247-L24-5-7618 119
GPRS/EGPRS Global Description Information
System

Standard BSC (no Standard BSC (full High capacity


GPRS) GPRS configuration) BSC

PPLD-9 PPCU-3 PPXU-3


PPLD-10
PPLD-11 PPXU-4
PPLD-12 PPCU-1
PPLD-13 PPXU-5
PPLD-14
PPCU-0

Tab. 6.3 Correspondence between the Boards of the Two Types of BSC

Since each PPXU is connected to the SNAP matrix by an 8 Mbit/s line, each PPXU
board, and as a consequence each PCU, is able to handle at most a data rate of 8
Mbit/s.
This data rate of 8 Mbit/s is split into 128 time slots of 64 kbit/s each. Since one of these
time slots is used to transmit the CRC related to the others, then 127 timeslots can be
used effectively.
This data flow is divided into two data rates:
1. a data rate constituted of 64 time slots of 64 kbit/s towards the Abis interface; this
flow allows the management of the Abis interface at most 64 X 4 = 256
GPRS/EGPRS channels (16 kbit/s each one), i.e., 256 PDTs
Please remember that, if either GPRS CS3 and CS4 coding schemes, or
i EGPRS coding schemes are used, 256 PDTs do not strictly correspond to
256 PDCHs.

2. a data rate constituted of 63 time slots of 64 kbit/s towards the Gb interface; this flow
allows the management of the Gb interface at most 63 Frame Relay Links (64 kbit/s
each one, see "7 Gb Interface").
Each PPXU board and, as a consequence, each PCU can handle up to 256 PDTs; to
reach 1280 PDTs (that is the number of packet switched resources provided by the high
capacity BSC), 5 boards (i.e., 1280/256 boards) have to be considered in service simul-
taneously; this means that 1+1 redundancy (used in the standard BSC) is no longer
possible and a different redundancy schema is provided.
The redundancy schema used is called “load balancing”: with this schema all six boards
are simultaneously in service and the packet switched traffic is distributed among all six
boards (see "8 Load Control for Packet Switched Services"); this implies that each
board will normally work in relax (the required real time traffic can be spread over 6
boards instead of 5).
When the BSC is fully equipped with six PCUs, it can handle up to 1536 PDTs (256 X
6) and 378 Frame Relay Links (63 X 6). If the 6th board is used for redundancy
purposes, the number of handled PDTs becomes 1280 (256 X 5).
With the high capacity BSC it is possible to configure up to 250 cells and, as a conse-
quence, up to 250 PTPPKF object instances.

120 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

6.1.3 High Capacity BSC with the New Rack


The current release foresees a new BSC rack to get a new BSC with high capacity.
Fig. 6.4 shows the high capacity BSC with the new rack.
The new rack contains the SNAP matrix and new PPXU, PPXL boards already present
in the high capacity BSC with the old rack; these boards have the same characteristics
already described in "6.1.2 High Capacity BSC with the Old Rack".
With respect to the old rack, the new BSC rack provides the following innovations:
• support to the new line interface card, the STLP, which is used in place of the QTLP
in the new rack; this includes the increase from 9 to 10 of the LICD number, and from
4 to 6 of the circuit number of LICD on which a PCM line can be created;
• support of 12 PCUs and PPXUs;
• new rack configuration and back plane: this will allow doubling the number of used
PPXU (PPXX supporting GPRS/EGPRS) from 6 to 12; it will also allow housing
10+2 STLP, i.e., one more card than QTLPs equipped in the old rack.
• new fan box for heat dissipation.
In order to achieve these goals, the following additional changes are introduced:
• the 2 EPWR (power supply devices in the expansion module) are removed to make
room for PPXU and STLP. For this reason, the PPXX and the STLP are provided with
an internal power supply, directly fed by the 48 V;
• on the top of the BSC, a box containing 6 fans has been introduced, powered by the
48 V, to cope with the increased heat generated by the BSC;
• sense points in the new board CPEX for monitoring the fan alarms have been intro-
duced.
As previously described (see "6.1.1 “Standard” BSC" and "6.1.2 High Capacity BSC
with the Old Rack"), the NTWCARD attribute allows specification of the BSC type. In fact
if NTWCARD is set to NTWSN16, the BSC is made by the old rack and it works with
SN16, PPCC, PPLD and PPCU boards; if the attribute value is NTWSNAP, then the old
rack is still used but in this case only the SNAP matrix and new PPXU and PPXL boards
are allowed.
If the user wants to use the new BSC rack, with new STLP boards (and obviously also
with SNAP and PPXU/PPXL boards), he must set the NTWCARD attribute equal to the
NTWSNAP_STLP value.
Regarding PPXU boards, the redundancy schema is always the “load balancing” one:
all the twelve boards are simultaneously in service and the packet switched traffic is
distributed among all the twelve boards (see "8 Load Control for Packet Switched
Services"); this implies that each board will normally work in relax (the required real time
traffic can be spread over 12 boards instead of 11).
When the BSC is fully equipped with twelve PCUs, it can handle up to 3072
GPRS/EGPRS PDTs (256 X 12) and 756 Frame Relay Links (63 X 12). If the 12th board
is used for redundancy purposes, the number of handled PDTs becomes 2816 (256 X
11).
Please remember that, if either GPRS CS3 and CS4 coding schemes, or EGPRS
i coding schemes are used, 756 PDTs do not strictly correspond to 756 PDCHs.

With the high capacity BSC it is possible to configure up to 400 cells and, as a conse-
quence, up to 400 PTPPKF object instances.

A30808-X3247-L24-5-7618 121
GPRS/EGPRS Global Description Information
System

Fig. 6.4 High Capacity BSC with the New Rack

6.1.4 PPCU and PPXU Redundancy and Configuration Rules


As previously described, with the standard BSC two boards are deputed to manage
GPRS service: they are named PPCU. For safety reasons both boards have a spare
copy. No dynamic data is present on the spare board so the redundancy schema is
defined, in BSC terminology, as cold standby. The creation of one PCU object implies
the consequent creation of the two related PPCU boards (active and standby).
When the high capacity BSCs are used, six or twelve boards are deputed to manage
GPRS and EGPRS services: they are named PPXU. In this case the 1+1 redundancy
is no longer possible and a different redundancy schema called “load balancing” is
provided (see 6.1.2 and 6.1.3)

122 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

It has been described that:


• the standard BSC can manage up to 150 GPRS cells, so the user can configure only
up to 150 PTPPKF object instances;
• the high capacity BSC with old rack can manage up to 250 GPRS/EGPRS cells, so
the user can configure up to 250 PTPPKF object instances;
• the high capacity BSC with old rack can manage up to 400 GPRS/EGPRS cells, so
the user can configure up to 400 PTPPKF object instances.
An important thing to be underlined (for all the BSC types) is that when the user config-
ures a cell to be used for packet switched services, i.e., when the user creates a
PTPPKF object instance, he does not have to assign the GPRS (or EGPRS) cell to a
specific PCU, but it is the system that assigns the cell dynamically to one of the available
PCUs.
This behavior, during PTPPKF creation, is a direct consequence of the load balancing
redundancy of the PPXUs:
– when a new cell is created, the system assigns it to a PCU (the less busy one);
– when a PCU becomes unavailable, the cells served by it are dynamically distributed
among the other PCUs.
This behavior also regards the PPCUs even if they use the cold stand-by redundancy.
In fact, using the standard BSC:
– when a new cell is created, the system assigns it to one of the two PCUs (the less
busy);
– when the active PPCU board fails, the stand-by one will replace the damaged one
(according to stand-by redundancy); if a couple of PPCU boards fails (i.e., if a PCU
becomes unavailable), all the GPRS/EGPRS traffic of the PCU will be managed by
the other couple of PPCU boards (i.e., by the other PCU) according to the load
balancing criteria (as it happens in the high capacity BSC).
The algorithm, that is used to distribute and redistribute the GPRS/EGPRS cells of one
BSC among the available PCUs, is described in "8 Load Control for Packet Switched
Services" chapter.

6.2 BTS Equipment Supporting GPRS and EGPRS


Regarding BTS equipment, different solution can be adopted according to the operator’s
needs. In fact, speaking about GPRS and EGPRS services, different kind of features are
provided according to the equipment used; the following solutions are available:
1. BTS equipment supporting EDGE and all GPRS coding schemes
2. BTS equipment supporting CS1 and CS2 coding schemes only
3. BTS equipment supporting GPRS CS3 and CS4 but not EDGE
EDGE support is limited to the BTSplus platform. To introduce EDGE into the network,
existing BTSplus sites must be upgraded with EDGE capable carrier units (E-CU)
featuring the new 8-PSK modulation technique. The E-CU hardware is able to handle:
– GSM, HSCSD and GPRS (with all its coding schemes) services;
– enhanced GPRS service (EGPRS).
To implement EGPRS, some GSM-CUs can be replaced by E-CUs at any arbitrary CU
rack position; mixed configurations with CUs and E-CUs as well as configurations with
E-CUs only are possible, but, as it has been said, the upgrade is only supported for
BTSE types belonging to the BTSplus generation.
Regarding GPRS coding schemes, the requirements are the following:

A30808-X3247-L24-5-7618 123
GPRS/EGPRS Global Description Information
System

• BTS1 base stations support GPRS CS3/CS4 coding schemes;


• BTS+, e-microBTS, picoBTS, support GPRS CS3/CS4 coding schemes (also if a
mixed configuration of GSM-CU and EDGE-CU boards is present);
Note that it is up to the operator to ensure the consistency between software configura-
i tion and BTS hardware.

6.3 PCU Frames and Dynamic Allocation on the Abis Interface


This chapter describes the distribution of GPRS/EGPRS traffic channels on the Abis
interface.
The Flexible Abis allocation strategy (Dynamic Allocation) is a general strategy used to
handle the Abis resources in a flexible way. A flexible Abis allocation strategy is neces-
sary to support GPRS CS3-CS4 coding schemes, and all EDGE coding schemes
requiring more than 16 kbit/s Abis throughput for specific radio channels.
With the dynamic Abis allocation, the number of Abis subslots that can be associated to
a radio timeslot depends on the service type.
From the hardware platforms point of view, the following equipment support the flexible
Abis allocation strategy:
• all the possible BSC configurations (standard BSC and high capacity BSCs,
see"6.1 Supported BSC Types");
• the BTSplus mainline with GSM-CU and EDGE-CU; picoBTS and enhanced micro
BTS; BTS1.
The introduction of CS3 and CS4 coding schemes for GPRS, and the EDGE feature
also have big impacts on the existing Abis interface, which must be modified:
• the Abis configuration of standard PCU frames (used for GPRS CS1/CS2 coding
schemes), which is based on a single 16 kbit/s slot, is not sufficient and does not
manage the transport of high data rates per air timeslot exploited by GPRS
CS3/CS4 coding schemes and EGPRS.
The EGPRS data is submitted via "concatenated PCU frames" (see 6.3.1), as well
as the GPRS data, in cases where the higher coding schemes (CS3/CS4) are
enabled. Hence the flexible Abis allocation strategy provides the opportunity to
assign to each air interface timeslot, from one to up to five 16 kbit/s Abis subslots
(16 kbit/s each one), in a flexible way;
• in cases where the capacity of each air interface timeslot can vary during runtime;
for GPRS CS3/CS4 or EGPRS, the flexible Abis allocation strategy adapts the Abis
capacity to the required air interface capacity (in cases of Link Adaptation/new TBF
establishments/old TBF releases). Note that the flexible Abis allocation strategy is a
slow process compared to GPRS/EGPRS Link Adaptation (see "10.5 Link Adapta-
tion"), hence the two processes must be synchronized;
• the total Abis capacity per BTS increases with the introduction of higher data rates
at the Um interface. Then, the flexible Abis allocation strategy must be coupled with
the management of up to 4 Abis PCM lines per BTS.
Tab. 6.4 shows how packet switched services can be mapped in 16 kbit/s, or N*16 kbit/s
Abis resources (per radio timeslot).

124 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

16 kbit/s N x 16 kbit/s

GPRS channels supporting EGPRS (up to 5x16 kbit/s)


CS1 and CS2 only GPRS channels supporting
CS3/CS4 (up to 2x16 kbit/s)

Tab. 6.4Mapping of Services onto Abis Resources

The flexible Abis allocation strategy coupled with the concept of concatenated PCU
frames gives the operator the following advantages:
– the Abis interface handling is more efficient: a common pool of Abis timeslots is
associated to a BTSM; then these Abis resources are shared between different
timeslots, carriers and even between different cells of the same base station site;
– EGPRS and GPRS Link Adaptation can be performed during runtime without loss
of service;
– unused capacity of an air interface timeslot can be released in the Abis interface and
exploited by other air interface timeslots;
– it is possible to reach a data rate up to about 60 kbit/s per packet data channel
(PDCH) on the Abis interface.
Generally speaking, the flexible Abis allocation strategy is managed by two different
processes:
1. the first task is the configuration one: the operator can assign to every BTSM a pool
of Abis timeslots. These timeslots will be used to transfer information between the
BTSM and the BSC;
2. the second task relies on the flexible allocation and release of resources taken from
the Abis pool. The Abis allocation algorithm is able to:
– assign sufficient Abis bandwidth to an air interface timeslot during run time;
– release bandwidth in case of congestion, according to service priorities and QoS
constraints.
The traditional Static Abis management is kept for backward compatibility with the
i previous releases, harmonizing the O&M management of “flexible” and “static” BTSM.

It must be clear that:


• flexible abis allocation means that the association between radio timeslots and
Abis timeslots is performed by radio signalling procedures. There is not a fixed
one-to-one (1 x 16 kbit/s) or one-to-two (2 x 16 kbit/s) association from air interface
timeslots to Abis subslots, in the BSC database;
• static abis allocation means that the association between radio timeslots and Abis
timeslots is performed during O&M procedures, stored into BSC database and
signalled to BTSM by O&M signalling procedures. The association is fixed during
runtime and can only be changed via O&M reconfiguration.
To simplify the configuration procedures, the operator commands used to configure both
“flexible” and “static” allocations for a BTSM are the same. In cases of “static” BTSM,
the static allocation between radio and Abis channels is performed by the system (BSC)
at configuration time.
In the following, the different topics related to this feature are discussed, considering:
– a discussion about concatenated PCU frames (see 6.3.1);
– hardware supporting flexible Abis allocation and concatenated PCU frames
(see 6.3.2);

A30808-X3247-L24-5-7618 125
GPRS/EGPRS Global Description Information
System

– configuration of the Abis interface (see 6.3.3);


– algorithms regarding flexible Abis allocation (see 6.3.4).

6.3.1 Concatenated PCU Frames


The PCU frame is used to transmit packet data traffic on the Abis interface, whereas to
transmit voice traffic, the similar TRAU frames are used.
Concatenated PCU can handle:
– EGPRS (MCS1,…., MCS9)
– GPRS Step II (CS3/CS4 data rates)
– GPRS Step I (CS1/CS2 data rates)
Concatenated PCU frames require partially more bandwidth (e.g., with 32 kbit/s, 48
kbit/s, 64 kbit/s and even 80 kbit/s) than the current 16 kbit/s ones. This higher band-
width is achieved by concatenating several 16 kbit/s subframes (1 up to maximum 5
EDGE subframes). In total, 5 subframes (instead of 4) are necessary for 60 kbit/s
MCS8/MCS9 because of the huge in band signaling overhead.
A PDCH with multiplexed EDGE and GPRS TBFs requires an Abis allocation due to the
highest used coding scheme. At the moment, a PDCH with a MCS9 TBF on it requires
16 kbit/s*5 = 80 kbit/s.
The subframes contain an index ranging from 1 to 5 called Sub-Frame-Counter (SFC)
plus several control parameters and spares, which can be used in the future.
The Sub-Frame-Counter (SFC) indicates the sub-frame number and provides the
sequence order of the 16 kbit/s channels. The SFC is coded by 5 bits, which allow a
maximum chain of up to 32 concatenated PCU subframes (see Fig. 6.5).
The receiving side (either PCU in uplink or BTS in downlink direction) is able to reas-
semble the subframes to achieve the original complete RLC/MAC block.

SFC=00000 SFC=00001 SFC=00010 SFC=000100

MCS/CS

Data Data

1st Subframe Following Subframes Last Subframe

Fig. 6.5 Fundamental Principle of Concatenated PCU Frames

Concatenated PCU frames transport, for each coding scheme of GPRS/EGPRS


services, the following number of bits, in the downlink and in the uplink direction respec-
tively (that represents the size of the transmitted RLC/MAC Block, see Tab. 4.2 and
Tab. 4.3):

126 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Coding Number of bits transmitted in DL/UL


scheme (corresponding to the total size of the
RLC/MAC block)

CS1 184
CS2 271
CS3 315
CS4 431
MCS1 209
MCS2 257
MCS3 329
MCS4 385
MCS5 478/487
MCS6 622/631
MCS7 940/946
MCS8 1132/1138
MCS9 1228/1234

Tab. 6.5

The useful payload part of the concatenated PCU frames is filled as follows:
• GPRS: Block Header, Data;
• EGPRS MCS1,...,6: Block Header, E, FBI/TI, Data
• EGPRS MCS7,...,9: Block Header, E, FBI/TI, Data 1st part, E, FBI/TI, Data 2nd part
Header Check Sequences (HCS), Block Check Sequences (BCS) and Tail Bits are
added by the BTS coder.
MSs using different coding schemes can be multiplexed on the same timeslots (PDCH)
on the air interface. Multiplexing of GPRS and EGPRS mobile stations is also possible
if concatenated PCU frames are used in both cases (i.e., on the same timeslot it is not
possible to multiplex users which are exploiting new concatenated PCU frames and
others working with the standard PCU frames).
The BTS and the BSC know how many Abis subslot are allocated to an air interface
channel and both know which PCU subframe with which SFC is mapped on each 16
kbit/s Abis subslot. That means: in cases of multiplexing several TBFs on the same
PDCH, for this PDCH, all TBFs have PCU frames with the same SFC on a specific Abis
subslot. Hence, due to the selected Coding Scheme, which is outlined in the control bits
of the first subframe, the mapping of the radio block payload to the PCU frame data bits
is given and it is also clear which PCU frame data bits must be filled with the pattern and
which (maybe) are idle.
The n*16 kbit/s subframes of an air interface timeslot are arbitrarily distributed over PCM
24/30 Abis lines: they are not necessarily allocated a block of subsequent Abis subslots,
which is of course possible. The subframes can be completely disordered on the PCM
lines of the BTSM as long as they are within the defined pool of the BTSM. They do not
have to guarantee any ordered sequence in ascending way due to increasing SFC.

A30808-X3247-L24-5-7618 127
GPRS/EGPRS Global Description Information
System

But, as it has been said, for a given PDCH, all allocated TBFs use the same Abis
subslots for concatenated PCU frames with the same SFC.
Although all subframes have an equal size of 40 Octets = 320 bit (16 kbit/s bit rate), the
shape of the first subframe and the other consecutive subframes is a little bit different.
Fig. 6.6 shows an example of the Abis mapping for a DL MCS9 radio block requiring 5
Abis subslots; the first subframe in Fig. 6.6 has a payload of maximum 216 bits, all
others can carry up to 272 bit. As soon as a selected coding scheme requires less than
the full number of data bits, the rest in the last data subframe are filled with a predefined
bit pattern, e.g., 11111111...... In cases of a coding scheme, which requires less
subframes than the PDCH has allocated, those completely unused subframes are idle
subframes also filled with the bit pattern 111111.... . These idle subframes are based on
the coding of the additional subframes.

Concatenated PCU Frames


SFC=00000 SFC=00001 SFC=00010 SFC=00011 SFC=00100
216 bits 272 bits

174 272 146 124 272 196


Data Data Data Data Data Data
Bits Bits Bits Bits Bits Bits

2 bits 2 bits 76 bits


40 Bits E, FBI E, FBI (11111..11)
RLC/MAC Filling Pattern
Header
(incl. USF)
1st RLC Data Block 2nd RLC Data Block

Fig. 6.6 Abis Mapping for a downlink MCS9 radio block requiring 5 Abis subslots

128 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Since users multiplexed on the same PDCH can not use a different number of PCU sub-
frames on Abis, idle PCU subframes with filling patterns are used on the Abis subslots
not carrying data payload, in order to extend all the concatenated PCU frames to the
same MCS-j (j=1,..., 9) configuration.
Let us consider an Abis channel that is allocated for a maximum bandwidth for a MS
using MCS9; in this case, MSs using MCSs lower than MCS9 have some idle PCU
frames with a filling pattern (e.g. 1111111...), due to the requirement that all TBFs on a
particular PDCH occupy the same Abis capacity, whether they need it or not.
Another case in which idle PCU-sub-frames are used to fill up the allocated Abis
capacity is when a Link Adaptation of a TBF to a lower data rates occurs (i.e.,
MCS9/MCS6, because of the impossibility of the air interface to maintain MCS9 with
good quality). The “unused” Abis capacity is filled with idle PCU sub-frames with filling
pattern, because to reduce signalling overhead, the release of allocated Abis capacity
is not executed immediately.
Standard PCU frames can be still used even combined with the flexible Abis allocation
i strategy; in fact dynamic Abis allocation does not imply the usage of concatenated PCU
frames. Standard PCU frames are used whenever the BTS does not support concate-
nated ones (see "6.3.2 Hardware supporting Flexible Abis Allocation and Concatenated
PCU Frames").

6.3.2 Hardware supporting Flexible Abis Allocation and Concatenated


PCU Frames
As it has been described, dynamic Abis allocation does not imply “concatenated PCU
frame” usage in packet flows.
Fig. 6.7 and Fig. 6.8 show the relationship among standard/concatenated PCU frames
and flexible/static Abis allocation, depending on the BTSE type and on the BSC type
(i.e., standard or high capacity BSC).

A30808-X3247-L24-5-7618 129
GPRS/EGPRS Global Description Information
System

High Capacity BSC


BTSplus, E-microBTS II
Both standard and
- Standard/Concatenated PCU frames supported
concatenated PCU frames
- CS1...CS4 supported Concatenated PCU frames are supported.
- MCS1...MCS9 supported on EDGE carriers
-Dynamic Abis allocation supported

BTS1 with BBSIG44, picoBTS, E-microBTS

- Standard/Concatenated PCU frames supported Concatenated PCU frames Dynamic Abis


- CS1...CS4 supported
- Dynamic Abis allocation supported
allocation

BTS1 without BBSIG44

- Only Standard PCU frames supported


- Only CS1 and CS2 supported
Standard PCU frames
- Dynamic Abis allocation supported

Fig. 6.7 High Capacity BSC: Relationship between PCU Frames and Abis Allocation according to the BTSE
Type

Standard BSC
BTSplus, E-microBTS II
Only standard
- Standard/Concatenated PCU frames supported
PCU frames
- CS1...CS4 supported Standard PCU frames are supported.
- MCS1...MCS9 supported on EDGE carriers
-Dynamic Abis allocation supported

BTS1 with BBSIG44, picoBTS, E-microBTS

- Standard/Concatenated PCU frames supported Standard PCU frames Dynamic Abis


- CS1...CS4 supported
- Dynamic Abis allocation supported
allocation

BTS1 without BBSIG44

- Only Standard PCU frames supported


- Only CS1 and CS2 supported
- Dynamic Abis allocation supported Standard PCU frames

Fig. 6.8 Standard BSC: Relationship between PCU Frames and Abis Allocation according to the BTSE Type

130 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

The following considerations can be done:


– standard PCU frames are used whenever GPRS CS3/CS4 and EDGE are not
supported;
– concatenated PCU frames are used whenever EDGE and/or CS3 and CS4 (more
than 16 kbit/s per radio timeslot) are supported;
– in all the previous cases, the mapping between radio timeslots and Abis timeslots is
dynamic (at channel activation);
– GPRS CS3/CS4 and EGPRS are not supported on “standard” BSC, due to its “low”
GPRS capacity (max. 128 PDCHs, decreasing to 64 PDCHs in case of CS3/CS4);
in this case, GPRS CS3/CS4 and EGPRS are not supported, and standard PCU
frames are always used.
The BSC software is backward compatible and it is able to handle BTSs running with old
software releases, supporting only static Abis allocation. Fig. 6.9 gives such an
example, when BTSs with old software releases are connected to a BSC with a release
supporting the flexible Abis allocation strategy. In this case only GPRS CS1/CS2 radio
channels are supported (GPRS CS3/CS4 or EGPRS capabilities cannot be configured).
There are not any problems handling such kind of situations since:
– static allocation and standard PCU frame format are implemented on BSC;
– operator commands for a release supporting the flexible Abis allocation strategy
have a “backward compatible” meaning and management (Abis pool definition is
internally handled in a “static” way for “old” BTS software releases);
– the BSC is able to reject operator commands not compatible with “old” BTS software
releases.

BSC Hardware

Software Release
supporting Dynamic
Allocation

All BTS Hardware with Dynamic Abis


Static Abis Software Release allocation

- Only Standard PCU frames supported Standard PCU frames


- Only CS1 and CS2 supported
Static Abis
- Only Static Abis allocation supported allocation

Fig. 6.9 BSC handling of BTS Equipment with Software Releases not supporting the Abis Dynamic Allocation

6.3.3 Configuration of the Abis Interface


As it has been described, the procedures used to configure both “flexible” and “static”
allocations for a BTSM, are the same. The only difference is that in cases of “static”
BTSM, the static allocation between radio and Abis channels is performed by the system

A30808-X3247-L24-5-7618 131
GPRS/EGPRS Global Description Information
System

(BSC) at configuration time. As a consequence, all the concepts explained below, are
valid, unless differently stated, for both “flexible” and “static” Abis allocations.
To manage the Abis allocations two concepts are introduced: the Abis Subpool and the
Abis Pool.
Referring to a specific BTSM, the Abis Subpool is a set of 16 kbit/s Abis subslots
belonging to a single PCMB line, routed together with the LPDLM instance (previously
associated to the BTSM) configured on the same PCMB line.
Remember that a BTSM can be connected to the BSC by, at most, four PCMB lines,
i and each line must contain at least one LPDLM related to the BTSM.

This is an operator constraint, valid for all kind of BSS configuration (star, loop, multidrop
with/without cross connections) and also for cross connectors external to the BSS
network elements. The “subpool” concept is necessary for O&M purposes, to manage a
correct fault propagation from LPDLM to Abis resources.
So, the user to connect the BSC to a specific BTSM can create a certain number of
subpools that will contain a specific number of timeslots of the Abis interface.
To configure an Abis subpool the SUBTSLB object is used. The SUBTSLB object
indicates one subslot of a PCMB line; when creating a SUBTSLB instance the user must
specify the following attributes:
– NAME: it indicates the subslots of a PCMB line, specifying the PCMB instance, the
slot [1..31] of the selected line, and the subslot number [0..3];
– ASSLAPD (Associated Lapd): it indicates the LPDLM instance (and as a conse-
quence the BTSM) that is related to this subslot.
So, to create on a PCMB line a subpool for a specific BTSM, the user must create more
instances of the SUBTSLB object, linking them to the same LPDLM instance (i.e., to the
same BTSM) by the ASSLAPD parameter.
Referring to a BTSM, the Abis Pool is the amount of 16 kbit/s Abis subslots reserved to
the BTSM for traffic services (i.e., it is the amount of SUBTSLB instances, configured on
different PCMB lines and associated, through the ASSLAPD, to the LPDLMs related to
the BTSM).
One more time it must be noted that:
– in cases of BTS supporting dynamic Abis allocation, Abis subslots are selected from
the Abis pool and allocated to radio channels at channel activation. In cases of
GPRS and EGPRS, changes of the Abis resources assigned to an air interface
timeslot are possible during TBF-operation via the channel modification command;
– in cases of static Abis allocation, Abis subslots are selected from the Abis Pool and
statically allocated to radio channels by O&M procedures; the relationship between
radio channels and Abis subslots is sent to the BTS by O&M Abis signalling (at radio
channel creation). The number of Abis subslots to be statically associated to the air
timeslot is always 1 for BTSs running with old SW releases.
Abis pools and subpools have the following properties and features:
• different Abis subpools, belonging to the same or different Abis pools, can be
defined on the same PCMB line;
• subpools can be distributed over all connected PCMB lines of a BTSM (at least one
subpool per line);
• the Abis subslots allocated to a radio channel may be distributed over different
subpools, over different PCM lines and it is not necessary at all to guarantee that the
subslots neighbor each other;

132 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

• overlaps between different pools and subpools are forbidden.


So, the PCU subframes belonging to a specific PDCH (or air interface timeslot) can be
distributed via all available Abis subpools, even if the subpools are located on different
PCMB lines.

6.3.4 Algorithms Regarding Flexible Abis Allocation


Regarding packet switched services, the BSC call processing (TDPC) handles dynam-
ically GPRS/EGPRS PDCHs which could require:
– 1 Abis subslot in case of CS1, CS2 (using standard PCU frames) and MCS1 coding
schemes;
– 2 Abis subslots in case of CS2 (using concatenated PCU frames) CS3/CS4 and
MCS2/MCS3/MCS4/MCS5 coding schemes;
– 3 Abis subslots in case of MCS6 coding scheme;
– 4 Abis subslots in case of MCS7 coding scheme;
– 5 Abis subslots in case of MCS8/MCS9 coding schemes.
Dynamic Abis allocation consists of the selection, for each radio channel of a cell, of one
or more idle and in service Abis subslots, belonging to the pool associated to the BTSM
that contains the cell; this selection is executed by the BSC during the channel activation
procedure; then the BSC informs the BTS by the CHANNEL ACTIVATION message.
Changes of the Abis resources assigned to a PDCH are also possible during TBF oper-
ations by the MODIFY ABIS CHANNEL message.
In cases of static Abis allocation, Abis subslots are selected from the pool and allocated
i to radio channels by O&M procedures; the radio channel/Abis subslot relationship is
sent to the BTS by O&M Abis signalling.

The Abis pools are present on TDPC database, related to the list of BTS (cells) fed by
the pool. Abis idle lists are built and updated according to the O&M operator commands
issued on the SUBTSLB object.
In cases of GPRS/EGPRS services, the number of Abis resources actually allocated at
service setup depends on several factors: required peak throughput, default applicable
coding scheme, Abis resources actual availability, PCU resources actual availability.
The 16 kbit/s Abis subslots, which are assigned to a Radio Channel (PDCH), can be
located arbitrarily at the Abis pool/subpools and must not obey any rules due to
increasing or decreasing subframe counter (SFC). The Abis subslots allocated to the
same radio channel may be distributed over different PCMB lines and it is not necessary
at all to guarantee that the subslots are adjacent to each other. As far as possible, the
Abis subslots for the same PDCH are selected from the same PCMB. For each allocated
Abis subslot, one PDT is allocated. But each Abis subslot of a Radio Channel is coupled
with a specific SFC, such that in cases of multiplexing several GPRS/EGPRS TBFs on
the same PDCH, the data of each TBF is transported in a fixed, predetermined way. All
PCU frames with the same SFC must be transported with the same 16 kbit/s Abis
subslot.
In cases of packet switched services, the initial Abis assignment can be changed
dynamically during operation due to:
– radio propagation conditions of the channels (Link Adaptation, see 10.5);
– pre-emption of circuit switched services over packet switched services (see
"5.3.6.2 Pre-emption of PDT Resources").

A30808-X3247-L24-5-7618 133
GPRS/EGPRS Global Description Information
System

The pool is managed with a “soft boundary policy”, which guarantees a minimum
percentage of Abis subslots for each cell. All the cells belonging to the same BTSM
share the same Abis pool; each cell may pick up Abis resources from the pool as long
as the ‘guaranteed minimum’ is left at the other cells’ disposal. The operator can set the
guaranteed minimum number of subslots per cell by the GUARMABIS parameter (BTS
object).
As it has been said, the BSC informs the BTS about the Air timeslots/Abis subslots rela-
tionship by two messages:
– the CHANNEL ACTIVATION message, when a new PDCH is set up;
– the MODIFY ABIS CHANNEL message, when for one or more already assigned
PDCHs a different number of Abis subslots is needed.
In the following sections the two case are discussed.

CHANNEL ACTIVATION Message


Both air interface (carrier, timeslot) and Abis resources (subslots from the Abis pool) are
included in the CHANNEL ACTIVATION message sent from the BSC to the BTS. After
having received the CHANNEL ACTIVATION message, the BTS connects the indicated
Abis resources with the Air interface timeslot.
The CHANNEL ACTIVATION message contains the following additional information:
• a list of 16 kbit/s Abis subslots assigned to the air interface timeslot; in cases of
multiple Abis subslots, information on the SFC numeration is given too;
• the PCU frame format type: it can be standard or concatenated.
If the CHANNEL ACTIVATION message activates a GPRS or EGPRS PDCH, the BTS
connects the concatenated PCU frames depending on their SFC with the particular Abis
channel(s). If the activation is successful, the BTS sends CHANNEL ACTIVATION ACK
to the BSC, otherwise it sends the CHANNEL ACTIVATION NACK.

MODIFY ABIS CHANNEL Message


If a GPRS/EGPRS channel changes its properties (e.g. according to link adaptation
procedure, the used coding scheme must be changed), the following sequence must be
respected in Abis Allocation and coding scheme change:
a) if there is a downgrading capacity, first the coding scheme of (all) TBFs on the PDCH
is adapted by PCU RLC/MAC signalling messages, then the Abis capacity is
changed by the MODIFY ABIS CHANNEL message; however that superfluous Abis
resources are not immediately released; unused PCU/Abis resources are released
after a given amount of time;
b) if there is an upgrading capacity, first the Abis capacity is changed by the MODIFY
ABIS CHANNEL message, then the Abis subslots are aligned and finally the coding
scheme of the TBF(s) can be switched. This process is possible only if enough Abis
capacity is free in the Abis pool and if enough PCU resources are available.
The MODIFY ABIS CHANNEL message must submit - just like the CHANNEL ACTIVA-
TION message - as a parameter the list of Abis subslots and their corresponding
subframe counters (SFC). It must be guaranteed that the lists within CHANNEL ACTI-
VATION and MODIFY ABIS CHANNEL are equal besides the changes which are made.
Furthermore, it must be clear that a MODIFY ABIS CHANNEL deletes the subframes
with the highest SFCs in cases of downgrading and adds subframes with adjacent
higher SFCs in case of upgrading. It is not allowed to modify the SFC to an already allo-
cated Abis subslot.

134 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

The MODIFY ABIS CHANNEL message contains the following information:


– the involved timeslot (specifying carrier and timeslot numbers);
– the new list of 16 kbit/s Abis subslots assigned to the air interface timeslot.

6.3.5 Abis over satellite links


The Abis interface is supported over satellite links. This transmission mode is the most
common implementation and it is often used to extend the GSM and GPRS/EGPRS
services to new locations with minimal infrastructure costs. For the reason that GSM
traffic grows also at remote sites, additional BTSs or a BSC may be deployed to support
higher traffic loads and/or a larger geographical area.
The satellite Abis configuration has the advantage that a minimal expense is requested
for deploying the service. An existing MSC and BSC can be used, which could possibly
support satellite connections to several remote locations.
The main disadvantage of the satellite Abis configuration is that the remote locations
relies heavily on the equipment located at the hub side so hand-offs and also eventual
subscriber to subscriber calls must go over the satellite link increasing load and traffic.
Besides the feature “AMR Link Adaptation” does not work together with the Abis Satel-
lite link.
The configuration’s parameters toghether with the related commands requested for the
Abis over satellite links are described in the manual: “CML:BSC”.
Besides the value of the attribute: “nRLCMAX” (this attribute determines the number of
the RLC data blocks before the Ack/Nack block is requested) of the PCU Managed
Object has been changed from “20” to “15” for reducing the problem related to the Abis
satellite’s delay. This attribute is not configurable.

6.4 Packet Switched Services Supported on CCCH/PCCCH


In the previous chapters (see 4.4.2) it has been described how it is possible to support
GPRS/EGPRS common signalling either on already existing CCCHs (shared CCCHs)
or on dedicated CCCHs (PCCCHs).
To avoid packet switched signalling load on traditional CCCHs, it is convenient to use
GPRS/EGPRS PCCCHs as soon as packet switched traffic increases beyond a certain
threshold, so that packet switched signalling traffic has no influence on normal signal-
ling, and the overall traffic capacity is improved.
These logical channels are mapped on different physical resources (see Fig. 6.10):
a) Dedicated CCCH: PCCCH is mapped in the multiframe of a Packet Data Channel
(PDCH); in this case, the common control signalling is carried in a logical channel
dedicated to GPRS/EGPRS traffic.
b) Shared CCCH: no dedicated control signalling channels exist for packet switched
data services, so that GPRS/EGPRS common control signalling packets access a
CCCH following its mapping rules. This mechanism is mandatory, whenever a dedi-
cated CCCH is not allocated.

A30808-X3247-L24-5-7618 135
GPRS/EGPRS Global Description Information
System

TDMA frame

CCCH PCCCH

PDCH

PCMB line

0 31

LAPD

Fig. 6.10 Mapping of CCCH/PCCCH Channels on the Abis Interface.

In both cases signalling messages are processed in the PCU, which is realized in BSC
by means of PPCU/PPXU cards (Peripheral Processors for GPRS/EGPRS).
In the following sections a short description is given about the message handling which
is implied by the described mechanisms (see Fig. 6.11):
a) Dedicated CCCH: messages are carried in a PCU frame on the 16 kbit/s timeslot
related to the physical PDCH, where the PCCCH is mapped. The timeslot is routed
via switching matrix directly to the PPCU/PPXU where the channel is processed.
b) Shared CCCH: messages are carried in the LAPD channel related to the BTSE. The
channel is routed via switching matrix to a PPLD where the LAPD protocol is
processed. The extracted messages are read by TDPC via Telephonic Bus from the
PPLD Dual Port RAM.
In the TDPC, the messages are analyzed: GPRS/EGPRS related messages are
written by TDPC via Telephonic Bus in the Dual Port RAM of the PPCU/PPXU,
where they are processed.

136 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Fig. 6.11 CCCH/PCCCH Message Handling.

The advantages of the first method (dedicated CCCH) are straightforward:


– on the air interface CCCH performances for normal GSM traffic are not reduced
because of the packet switched data messaging;
– on the Abis interface the capacity of the LAPD link is not shared between GSM and
GPRS/EGPRS traffic;
– the TDPC does not waste real time to route packet switched data messages toward
PPCUs/PPXUs and to multiplex in LAPDs the messages received from the PPCUs
– the Telephonic Bus is not loaded (twice) by the exchange of messages among
PPLD, PPCU/PPXU and TDPC.
On the other side shared CCCHs is supported in any case to provide the first access
when no specific GPRS/EGPRS signalling channels are allocated.
Shared CCCHs are the only way to allow Class B MSs (see "9.1 Mobile Stations for
Packet Switched Services") attached to GPRS/EGPRS to listen to their circuit switched
paging channel on CCCH, when the optional Gs interface between the MSC and the
SGSN is not implemented (see "9.8.3.1 Network Operation Modes for Paging").

A30808-X3247-L24-5-7618 137
GPRS/EGPRS Global Description Information
System

7 Gb Interface
The Gb interface connects the BSC to the SGSN, transferring signalling information and
user data. Several BSCs may be interfaced to one SGSN on the Gb interface.
The main characteristics of the Gb interface are:
a) The resources are given to a user upon activity (when data is sent or received) and
they are reallocated immediately thereafter; this is in contrast to the A interface,
where a single user has the only use of a dedicated physical resource throughout
the lifetime of a call, irrespective of activity;
b) GPRS/EGPRS signalling and user data are sent in the same physical channel. No
dedicated physical resources are required to be allocated for signalling purposes
(like e.g., the A interface where SS7 links are used to transmit signalling between
the BSC and the MSC).
c) The Gb interface is supported also over satellite links. Specific procedures for
defining the multiple frame operation mode necessary for the transmission of the “I”
frames have been implemented. The “T200” and “k” attributes are not configurable
by the user because they are not used. The transmission of information at applica-
tion level is supported by UI frames (Unacknowledged information/Downlink-Unit-
data). As a consequence of this the matter than the Gb interface can be suppported
also by satellite links is not relevant.In any case only the parameters defined in the
specification: “Q933” can be configurable, but their values do not depend from data
link/physical layer over satellite.
The protocol stack of the “Gb” interface is illustrated in the Fig. 7.1.

Fig. 7.1 Gb Interface: Protocol Stack

The several layers realize the following functions:


• L1: it specifies the Layer 1 of the Gb interface. Frame Relay (FR) is used for
GPRS/EGPRS in a first phase.
• Network Service (NS): it performs transport of NS Service Data Units (SDU)
between the SGSN and BSS. The Gb interface is based on Frame Relay (FR) as
specified in GSM 08.16.

138 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

FR supports high data rate transmission with low delay. Frames of different sizes
may be transmitted. FR performs congestion control and error detection, however
error correction is not supported.
• BSSGP: the primary functions of the Base Station Subsystem GPRS protocol
(BSSGP) are:
– providing connection-less links between the SGSN and the BSS (layer 2 level);
– providing tools for bi-directional control of data flow;
– handling paging requests from the SGSN to the BSS.
• LLC (Logical Link Control layer): provides logical links between a MS and the corre-
sponding SGSN. The transport of both data and signalling is supported;
• SNDCP (SubNetwork Dependent Convergence Protocol): supports a direct peer to
peer (i.e.,, point-to-point) communication between a MS and a SGSN. User data is
transported from a network layer protocol, e.g. IP or X.25.
The NS layer of the Gb interface is split into a Network Service Control part and a Sub
Network Service part. The Service Control part is independent from the physical realiza-
tion of the network, whereas the Sub-Network Service entity is the Frame Relay
protocol.

7.1 Physical Layer


Four types of configurations are possible to connect the BSC to the SGSN:
1. a direct line (e.g., PCM30, PCM24) between the two entities (static and permanent
physical point to point connections);
2. an intermediate frame relay network;
3. Nailed Up Connection (NUC) through the MSC via a frame relay network;
4. NUC through MSC, without using an intermediate frame relay network.
The different configurations are illustrated in Fig. 7.2.

A30808-X3247-L24-5-7618 139
GPRS/EGPRS Global Description Information
System

Fig. 7.2 Different Connection Types between the BSC and the SGSN.

The Gb interface is realized by PCM lines:


• in cases of direct connections between the BSC and the SGSN, PCM lines are
called PCMG;
• in cases of connections through the MSC (and TRAU), PCMA lines are used.
The PCMG object represents the PCM line used to connect the BSC and the SGSN,
without passing through the MSC.

Functional object Meaning

PCMG Represents the direct physical connection


between BSC and SGSN.

Tab. 7.1 PCMG Object

On the PCMG line, 31 physical channels, of 64 kbit/s each one, can be handled (slot 0
is always use for synchronization purposes).
In case of standard BSC (see 6.1.1), up to two PCMG lines can be configured:
– PCMG:0;
– PCMG:1.
In fact in this case two PCMG lines are enough to handle the 32 X 64 kbit/s channels
(16 channels for each PCU) that can be equipped toward the Gb interface, also
providing the possibility to have fault redundancy.
When the high capacity BSC with the old rack is used (see 6.1.2), in order to completely
exploit the bandwidth that the 6 PPXUs offer toward the Gb interface (in total 378 time
slots at 64 kbit/s), an increase of the PCMG number is necessary. For E1 lines (31 time
slots), 12 lines are enough, while for the T1 lines (PCM24 mode), 16 PCMG lines are
necessary: so this is the number of PCMG that is possible to configure at most with this
kind of BSC.
When the high capacity BSC with the new rack is used (see 6.1.3), in order to completely
exploit the bandwidth that 12 PPXUs offer toward the Gb interface (in total 756 time slots
at 64 kbit/s), an increase of the PCMG number is necessary. For E1 lines (31 time slots),
24 lines are enough, while for the T1 lines (PCM24 mode), 32 PCMG lines are neces-
sary: so this is the number of PCMG that is possible to configure at most when the new
BSC rack is used.
As it has been described in "6 Hardware and Software Architecture", each PCU
manages the packet switched data traffic of a specific number of cells; to transmit packet
data (or signalling) related to these cells, each PCU can use all the PCMG lines config-
ured for the BSC. In other words, the PCM line is not statically assigned to one PCU, but
to the whole BSC.
This line can be connected in one circuit of LICD without any restrictions. The LICD
circuit using QTLP V2 can be programmed in transparent mode and in this way we can
connect 2 PCM lines to 1 LICD circuit.
The following attributes are involved in PCMG configuration:

140 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

• PCML: this attribute identifies the LICD number (range 0 to 9), the CIRCUIT number
(range 0 to 5) and the TRUNK (A or B) to which the PCM line is connected;
The range 0..5 of the CIRCUIT number is valid when STLP boards are used
i in the BSC (i.e.,, when the new BSC rack is used), otherwise that admitted
range is 0..3.

• CRC: this attribute indicates if CRC-4 signal handling for PCM 30 line or CRC-6
signal handling for PCM 24 line is Enabled on PCMG line;
• CODE: this attribute selects the line transmission code to be provided on the line;
• NUA: this attribute enables or disables handling of not urgent alarms on PCMG line;
• BER (Bit Error Rate): this attribute indicates the threshold that, if exceeded, the line
must be put in Disabled state;
• BAF: this attribute defines frame alignment bits that can be set by the operator;
• LOWBER (Lower Bit Error Rate): this attribute is relevant only for PCM24 lines;
• REMAL (Remote AlarmType): this attribute is relevant only for PCM24 lines.

The Gb interface physical layer is specified in GSM 08.14; it is called Frame Relay Link
(FRL).
The Frame Relay Link is a n X 64 kbit/s physical channel, created over a PCM line.
These physical channels can be created grouping either neighboring or spaced time
slots of the PCM line; more than one physical channel can be created over a single line
(see Fig. 7.3).

FRL_1 (Channelized FRL))


PCM line

0 31

PCM line

0 31
FRL_2 FRL_3
(Fractional FRL) (Fractional FRL)

Fig. 7.3 Example of Frame Relay Links

In case of direct connections between the BSC and the SGSN, frame relay links are
created over PCMG lines, whereas in case of connections through the MSC, the FR
links are created over PCMA lines.
The FRL object represents the physical channel over the Gb interface between the
BSC and the SGSN.

A30808-X3247-L24-5-7618 141
GPRS/EGPRS Global Description Information
System

Functional object Meaning

FRL Represents the physical link connection on


Gb interface.

Tab. 7.2 FRL Object

In case of A interface connections, the 64 kbit/s time slots are reserved on PCMS (and
PCMA) lines and handled in TRAU as transparent channel. In case of direct Gb interface
connections (i.e., connections built without passing through the MSC), PCMG lines are
dedicated to SGSN connection, and the FRL occupies one or more 64 kbit/s timeslots.
The choice between direct connections or A interface connections can be done in base
of the bandwidth required on Gb interface (in case of a small number of FRL links, it is
advantageous to use A interface connections).
In case of A interface connections, with multislot links, the customer must guarantee that
the MSC is able to ensure the sequence. If the MSC is not able to guarantee this feature,
only single timeslot frame relay links can be configured.
When a standard BSC is used (see 6.1.1), up to 32 frame relay links can be created for
each BSC (with range 0 to 31). As described in "6 Hardware and Software Architecture",
each PCU is able to handle 1 Mbit/s data flow towards the Gb interface. This flow corre-
sponds to a flow obtained by 16 slots (64 kbit/s each one) on a PCM line. This factor
determines the maximum number of Frame Relay links that can be configured for each
PCU, and the capacity in terms of bit/rate; in fact for each PCU:
– up to 16 FRLs of 64 kbit/s can be configured;
– or only a single FRL with 1Mbit/s can be configured.
When the high capacity BSC with the old rack is used (see 6.1.2), up to 378 frame relay
links can be created for each BSC (with range 0 to 377). As described in "6 Hardware
and Software Architecture", each PCU is able to handle a 4 Mbit/s data flow towards the
Gb interface. This flow corresponds to a flow obtained by 63 slots (64 kbit/s each one)
on a PCM line. This factor determines the maximum number of Frame Relay links that
can be configured for each PCU, and the capacity in terms of bit/rate; in fact for each
PCU at most 63 FRLs of 64 kbit/s can be configured.
When the high capacity BSC with the new rack is used (see 6.1.3), up to 756 frame relay
links can be created for each BSC (with range 0 to 755). As described in "6 Hardware
and Software Architecture", each PCU is able to handle a 4 Mbit/s data flow towards the
Gb interface. This flow corresponds to a flow obtained by 63 slots (64 kbit/s each one)
on a PCM line. This factor determines the maximum number of Frame Relay links that
can be configured for each PCU, and the capacity in terms of bit/rate; in fact for each
PCU at most 63 FRLs of 64 kbit/s can be configured.
When creating a Frame Relay Link the operator specifies which PCU it belongs to, using
the PCUID attribute. This attribute indicates the pathname of the PCU managing the
FRL.
The operator indicates:
1. the PCM line on which the link is created, using the GLK attribute;
2. the number of slots that constitutes the FRL, using the GTS attribute.
For example:
– setting GTS= 3, allows configuration of a 64 kbit/s Frame Relay link on the slot
number 3 of the PCM line which is specified by the GLK attribute (see Fig. 7.4);

142 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

– setting GTS= 3&4&5&6, allows configuration of a 256 kbit/s Frame Relay link on
slots number 3, 4, 5 and 6 of the PCM line which is specified by the GLK attribute
(see Fig. 7.5);
– setting GTS= 3&4&7&8, allows configuration of a 256 kbit/s Frame Relay link on
slots number 3, 4, 7 and 8 of the PCM line which is specified by the GLK attribute
(see Fig. 7.6).

0 31

64 kbit/s Frame Relay Link

Fig. 7.4 Example of Frame Relay Link (GTS=3).

0 31

256 kbit/s Frame Relay Link

Fig. 7.5 Example of Frame Relay Link (GTS=3&4&5&6).

0 31

256 kbit/s Frame Relay Link

Fig. 7.6 Example of Frame Relay Link (GTS=3&4&7&8).


The operator, by the FRSTD attribute, can also optionally indicate the frame relay stan-
dard to be used (regarding the frame relay structure, see 7.2.1.2).
Supposing to configure, for each PCU, 2 FRLs, these links can be distributed on the Gb
interface in different manners, by setting the GTS attribute, e.g.:
– it is possible to put the two links on the same PCMG line;
– it is possible to distribute them on two different PCMG lines (this situation is obvi-
ously better than the previous one, since the redundancy of the links is provided; in
fact in case of fault of one PCMG line, the other one allows the connection between
the BSC and the SGSN to be maintained);
– it is possible to put one of them on one PCMG line, and the remaining one on one
PCMA line;
– it is possible to put both of the links over PCMA lines.

A30808-X3247-L24-5-7618 143
GPRS/EGPRS Global Description Information
System

When the links are created over different PCMA lines, and these lines belong to the
same TRAU module (i.e., the lines correspond to the same PCMS line), the FR links
must have different timeslot values for the GTS attribute.
Instead, if the lines belong to different TRAU modules this problem does not exist. This
last solution is obviously better than the previous one, since it provides the redundancy
of FRLs.
Remember that the PCMG/PCMA lines are shared between the configured PCUs,
i whereas each Frame Relay Link is associated to a specific PCU according to the PCUID
value.

7.2 Network Service Layer


The Network Service layer provides a reliable connection between the BSC and the
SGSN; this reliable connection is realized:
a) within the FR network, when such network exists between the two entities;
b) with a direct link, in cases of point-to-point connections.
Error detection is performed, while error recovery is left to upper layers.
The Network Service entity is composed of (see Fig. 7.7):
1. the Sub-Network Service (i.e., the Frame Relay protocol), which is an entity depen-
dent on the intermediate Gb interface network;
2. the Network Service Control, i.e., a control entity independent from that network.

Fig. 7.7 Network Service Layer

7.2.1 Sub-Network Service: Frame Relay on Gb Interface


On the Gb interface, and specifically inside each Frame Relay physical link, only Perma-
nent Virtual Circuits are implemented.
A Permanent Virtual Circuit (PVC) is an end-to-end logical communication link between
the BSS/PCU and the SGSN, irrespective of the exact configuration of the Gb interface.
These PVCs are created inside the FR physical links, and each FRL can contain more
than one PVC.
For PVCs there is no call set-up or clearing: a connection to the frame relaying node
must be in place from the configuration point of view.

144 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

The NSVC (Network Service Virtual Connection) object represents the end-to-end
permanent virtual connection between the BSC and the SGSN.

Functional object Meaning

NSVC Represents the end-to-end communication


between BSS and SGSN.

Tab. 7.3 NSVC Object

Each NSVC is identified by the Network Service Virtual Connection Identifier (NSVCI).
Up to 65536 NSVCIs can be created between a BSC and the SGSN. For each FRL (i.e.,
for each Frame Relay physical link) more than one NSVC can be created.
Referring to Fig. 7.8 there is a set of principles that apply to the Gb FR network:
• the physical link is the Frame Relay bearer channel (allocated timeslots in a PCMG
or a PCMA line);
• the NSVC is the FR PVC;
• the FR PVC (NSVC) provides an end-to-end connection through the FR network.
The Network Service Virtual Link (NSVL) is the local link in one end of the FR PVC,
i.e it is the link at the User Network Interface (UNI);
• the Data Link Connection (DLC) defines the entry point to the FR network. A DLC is
identified by a DLC Identifier (DLCI);
• the Network Service Virtual Link Identifier (NSVLI) is the DLCI together with the
bearer channel identifier (FRL). A physical link supports one or more NSVLs; each
one is identified by a NSVLI.

Frame Relay physical link

Fig. 7.8 Gb Interface with a Frame Relay Network

A30808-X3247-L24-5-7618 145
GPRS/EGPRS Global Description Information
System

When creating a new PVC, i.e., when creating a new instance of the NSVC object, the
user must specify the following:
1. the Network Service Virtual Connection Identifier (NSVCI) of the NSVC, i.e., the
common and absolute identification of the virtual connection between the SGSN and
the BSS; to specify this value he uses the NSVCI parameter;
2. the Network Service Virtual Link Identifier (NSVLI) to identify the NSVC on the local
(BSS) side. To specify this value he uses the NSVLI parameter; this parameter is
composed of two fields:
– the first one (FRLN) indicates the Frame Relay physical link on which the perma-
nent virtual circuit is created;
– the second one (DLCIN) indicates the DLCI number; this identifier (that is the
address of Frame Relay packets, see "7.2.1.2 Frame Relay Structure") allows a
distinction between different NSVCs that belong to the same physical Frame
Relay link.
The mapping of the DLCI parameter is as follows:

DLCI value Mapping


0 In band signalling
1-15 Reserved
16-511 Available for user information
512-991 Available for user information

Since Frame Relay Physical links are statically associated to a single PCU, even the
i NSVCs created inside this FRL are handled by a single PCU. The PCU will then share
its traffic among all its NSVCs.
So, each PCU can manage:
- a set of frame relay physical links (FRLs);
- a set of NSVCs, for each FRL.
- NSVCs belonging to different FRLs are distinguished by the FRLN attribute;
- NSVCs belonging to the same FRL are distinguished by the DLCIN attribute.

All the NSVCs configured for a PCU constitute the so called NSVC group; this group is
i identified by the Network Service Entity Identifier (NSEI).
The NSEI is the logical entity of the SGSN that manages a single PCU; as a conse-
quence it identifies, besides the PCU, all the NSVCs configured for the Packet Control
Unit.
The NSEI value, that identifies the PCU and its NSVCs is configured by the NSEI
parameter.

If a direct end-to-end PCMG line connection is used between the BSC and the SGSN
i (i.e., if a Frame Relay Network is not used), the two values related to one NSVC are the
same; i.e., the NSVLI value at the BSS side is equal to the NSVLI value at the SGSN
side.
When an intermediate FR network is used in connecting the BSS and the SGSN, the
NSVLI values, of the same NSVC, can have a different value at the SGSN side and at
the BSS side.

146 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

7.2.1.1 Examples of Addressing


In order to provide end-to-end communication between the SGSN and the BSS irre-
spective of the exact configuration of the Gb interface, the concept of Network Service
Virtual Connection (NSVC) is used. At each side of the Gb interface there is a one-to-
one correspondence between NSVCs and NSVLs.
The creation of a NSVC may be as follows (see Fig. 7.9):

Database object instance: NSVC-3 Defining the end-to-end connec-


tion between the SGSN and the
BSC/PCU.
Parameters for NSVC-3: NSVCI=5 Internal identifier of the FR
network connecting each side of
the network.
NSVLI=0-111 FRL object 0, DLCI 111.
Local connection at BSC/PCU
side.
Similarly a NSVLI value must be
defined at the SGSN side.

Fig. 7.9 Creation of a NSVC

To well understand previous concepts, some examples regarding the configuration of


both Frame Relay Links and Permanent Virtual Connections (NSVCs) are described.
Let’s consider a BSC that is connected to the SGSN by two direct PCMG lines (PCMG-
0 and PCMG-1).

EXAMPLE 1: BSC Configured with One PCU and Two Frame Relay Links of 64
kbit/s each.
Two frame relay links of 64 kbit/s each have been created for a BSC configured with a
single PCU. The PCU has been configured with a NSEI value equal to 2354
(see Fig. 7.10).
The PCU sees a total bandwidth of 128 kbit/s (64 kbit/s + 64 kbit/s).

A30808-X3247-L24-5-7618 147
GPRS/EGPRS Global Description Information
System

FRL:0
NSEI = 2354
PCUID:PCU-0
GLK:PCMG-0
GTS:2

PCMG-0
PCU- 0
0 31

PCMG-1

0 31

FRL:1
PCUID:PCU-0
GLK:PCMG-1
GTS:5

Fig. 7.10 BSC Configured with One PCU and Two FR Links (64 kbit/s each).
Supposing now to create a PVC for each FRL; Tab. 7.4 shows possible values that can
be used to create the two virtual connections. As it can be seen, DLCI values of the two
created NSVCs can be equal, since the two NSVCs belong to two different FRLs.

NSVC belonging to FRL:0

NSVCI 494
FRLN 0
NSVLI
DLCI 100
NSVC belonging to FRL:1
NSVC 512
FRLN 1
NSVLI
DLCI 100

Tab. 7.4 Example of Setting of NSVC Values.

EXAMPLE 2: BSC Configured with One PCU and Two Frame Relay Links of 128
kbit/s each.
Two frame relay links of 128 kbit/s each have been created for a BSC configured with a
single PCU. The PCU has been configured with a NSEI value equal to 2354
(see Fig. 7.11).
The PCU sees a total bandwidth of 256 kbit/s (128 kbit/s + 128 kbit/s).

148 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

FRL:0
NSEI = 2354 PCUID:PCU-0
GLK:PCMG-0
GTS:2&3

PCMG-0
PCU- 0
0 31

PCMG-1

0 31

FRL:1
PCUID:PCU-0
GLK:PCMG-1
GTS:5&8

Fig. 7.11 BSC Configured with One PCU and Two FR Links (128 kbit/s each one).
Supposing now to create a PVC for each FRL; Tab. 7.4 shows possible values that can
be used to create the two virtual connections.
It can be seen, that from the NSVC configuration point of view, there isn’t any difference
with respect to the previous example, even if the FRL:1 has been created using two non-
adjacent timeslots.
Obviously the network must be enable to support one FRL created with two non-neigh-
boring slots.

EXAMPLE 3: BSC Configured with Two PCUs and Two Frame Relay Links of 128
kbit/s each.
In this case, the BSC contains two PCUs. The PCU-0 has been configured with a NSEI
value equal to 2354, while the PCU-1 is identified by the NSEI= 7564 (see Fig. 7.12).
For each PCU, two frame relay links of 128 kbit/s each have been created; the PCU sees
a total bandwidth of 256 kbit/s (128 kbit/s + 128 kbit/s).

A30808-X3247-L24-5-7618 149
GPRS/EGPRS Global Description Information
System

FRL:0 FRL:2
NSEI = 2354 PCUID:PCU-0 PCUID:PCU-1
GLK:PCMG-0 GLK:PCMG-0
GTS:2&3 GTS:8&9

PCMG-0
PCU- 0
0 31
PCU- 1

PCMG-1

0 31

FRL:1 FRL:3
NSEI = 7564 PCUID:PCU-0 PCUID:PCU-1
GLK:PCMG-1 GLK:PCMG-1
GTS:5&7 GTS:10&11

Fig. 7.12 BSC Configured with Two PCUs and Two FR Links each one.
Supposing now to create a PVC for each FRL; Tab. 7.5 shows possible values that can
be used to create the two virtual connections for the PCU-0, and possible values that
can be used to create the two virtual connections for the PCU-1.
The NSEI identifier of the PCU-0, not only identifies the PCU, but also NSVCs used to
support the traffic of the PCU-0; in the same way the NSEI identifier of the PCU-1, not
only identifies the PCU-1, but also NSVCs used to support its traffic.

NSVC belonging to FRL:0


NSVCI 480
FRLN 0
NSVLI
DLCI 163
NSVC belonging to FRL:1
NSVCI 555
FRLN 1
NSVLI
DLCI 100
NSVC belonging to FRL:2
NSVCI 574
FRLN 2
NSVLI
DLCI 100

Tab. 7.5 Example of Setting of NSVC Values for both PCU-0 and PCU-1

150 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

NSVC belonging to FRL:3


NSVCI 575
FRLN 3
NSVLI
DLCI 216

Tab. 7.5 Example of Setting of NSVC Values for both PCU-0 and PCU-1

7.2.1.2 Frame Relay Structure


Core functions of the NS Sub Network Service provide necessary data link functions to
permit routing and relaying, but excludes those associated with sequencing, most forms
of error detection, error recovery and flow control.
Referring to the Frame Relay frame format (see Fig. 7.14), the Sub Network Service
functionality provides for:
a) Delimiting, alignment and transparency using the “Flag” field.
b) Multiplexing/De-multiplexing using the “Address” field. This function permits one
or more core connections to exist across a single physical connection.
c) Error detection using the “FCS” field (no Error Recovery). Frame Relay uses a
common error-checking mechanism known as the cyclic redundancy check (CRC).
The CRC compares two calculated values to determine whether errors occurred
during the transmission from source to destination. Frame Relay reduces network
overhead by implementing error checking rather than error correction. Frame Relay
typically is implemented on reliable network media, so data integrity is not sacrificed
because error correction can be left to higher-layer protocols running on top of
Frame Relay.
d) Congestion control using the “FECN”, “BECN”, and ”DE” fields. This function
permits core entities to detect congestion, to optionally notify peer entities of
congestion conditions, and to discard data units in response to congestion.
Frame Relay implements two congestion-notification mechanisms:
– Forward-explicit congestion notification (FECN);
– Backward-explicit congestion notification (BECN)
FECN and BECN features are controlled by a single bit contained in the Frame
Relay frame header. The Frame Relay frame header also contains a Discard Eligi-
bility (DE) bit, which is used to identify less important traffic that can be dropped
during periods of congestion.
The FECN mechanism is initiated when a DTE device (e.g., in our case the SGSN)
sends Frame Relay frames into the network (see Fig. 7.13).

A30808-X3247-L24-5-7618 151
GPRS/EGPRS Global Description Information
System

Fig. 7.13 Frame Relay Network Connecting two DTE Devices

If the network is congested, DCE devices (switches) set the value of the frames’
FECN bit to 1. When the frames reach the destination DTE device, the Address field
(with the FECN bit set) indicates that the frame experienced congestion in the path
from source to destination. The DTE device can relay this information to a higher-
layer protocol for processing. Depending on the implementation, flow-control may be
initiated, or the indication may be ignored.
DCE devices set the value of the BECN bit to 1 in frames travelling in the opposite
direction of frames with their FECN bit set. This informs the receiving DTE device
that a particular path through the network is congested. The DTE device can then
relay this information to a higher-layer protocol for processing. Depending on the
implementation, flow-control may be initiated, or the indication may be ignored.
The Discard Eligibility (DE) bit is used to indicate that a frame has lower importance
than other frames. DTE devices can set the value of the DE bit of a frame to 1 to
indicate that the frame has lower importance than other frames. When the network
becomes congested, DCE devices will discard frames with the DE bit set, before
discarding those that do not. This reduces the likelihood of critical data being
dropped by Frame Relay DCE devices during periods of congestion.
Two parameters are involved in the congestion control procedure:
– TCONG: this parameter allows the user to configure the width of the observation
window used for congestion detection. The congestion detection regards the path
from the SGSN to the BSC (i.e., it regards the frame relay frames sent by the
SGSN to the BSS).
If, during the time defined by TCONG, the number of frames indicating congestion
is equal or greater than the number of frames indicating no congestion, the
congestion state is notified to upper layers;
– TCONOFF: after a congestion notification to upper layers, no other notifications
are foreseen for a length of time defined by TCONOFF. This timer is needed to
provide a hysteresis time in order to ensure that the traffic reduction at mobile
station can be effective.

152 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

All data link peer to peer communications use frames conforming to the format shown
in Fig. 7.14.

Fig. 7.14 Frame Relay Frame Structure

The basic Frame Relay fields are the following:


• Flags: delimits the beginning and end of the frame. The value of this field is always
the same and is represented either as the hexadecimal number 7E or the binary
number 01111110.
• Address: contains the following information:
– DLCI: the 10-bit DLCI is the essence of the Frame Relay header. This value repre-
sents the virtual connection between the DTE device and the switch. Each virtual
connection that is multiplexed onto the physical channel will be represented by a
unique DLCI. The DLCI values have local significance only, which means that
they are unique only to the physical channel on which they reside. Therefore,
devices at opposite ends of a connection can use different DLCI values to refer to
the same virtual connection. Since there is neither D-channel nor layer 2 manage-
ment functionality, the available DLCI values usable for user information are
almost the whole DLCI range but DLCI 0, which is reserved for layer 3 message
transfer (STATUS and STATUS_ENQUIRY, see "7.2.1.3 Procedures for PVCs");
– Extended Address (EA): it is used to indicate whether the byte in which the EA
value is 1 is the last addressing field. If the value is 1, then the current byte is
determined to be the last DLCI octet. Although current Frame Relay implementa-
tions use a two-octet DLCI, this capability allows for longer DLCIs to be used in
the future. The eighth bit of each byte of the Address field is used to indicate the
EA.
– C/R: it is the bit that follows the most significant DLCI byte in the Address field.
The C/R bit is not currently defined.
– Congestion Control: it consists of the three bits that control the Frame Relay

A30808-X3247-L24-5-7618 153
GPRS/EGPRS Global Description Information
System

– congestion-notification mechanisms. These are the FECN, BECN, and DE bits,


which are the last three bits in the Address field.
• Data: contains encapsulated upper-layer data. Each frame in this variable-length
field includes a user data or payload field that will vary in length up to 16,000 octets.
This field serves to transport higher-layer protocol packets (PDUs) through a Frame
Relay network.
• Frame Check Sequence: ensures the integrity of transmitted data. This value is
computed by the source device and verified by the receiver to ensure integrity of
transmission.

7.2.1.3 Procedures for PVCs


For each group of PVCs belonging to a FRL, a periodic polling procedure is used in
acquiring general status about the connection between the BSC and the SGSN. The
polling interval is defined by the T391 timer. Every T391 seconds the PCU sends a
STATUS ENQUIRY message to the network to retrieve some information (optionally the
network may initiate the polling procedure).
The information regards:
a) notification of the addition of a PVC: used to notify users of newly added permanent
virtual circuits;
b) detection of the deletion of a PVC: used to notify users of deleted permanent virtual
circuits;
c) notification of the availability (active) or unavailability (inactive) state of a configured
PVC: used to determine changes in status of configured PVCs;
d) link integrity verification: used in determining the in-channel signalling link DLCI-0.
Establishing and releasing a logical connection is accomplished by exchanging
messages via DLCI-0. The Link Integrity verification procedure is required since
DLCI-0 contains unnumbered information (UI) frames at Level 2.
The periodic polling procedure allows the PCU to retrieve the previous information; the
procedure is shown in Fig. 7.15.:
1. every T391 seconds, the PCU sends a STATUS_ENQUIRY message to the network;
2. the network answers with a STATUS message; two types of STATUS messages can
be used:
– if no PVCs have been added to or deleted from the FRL, or if the state of config-
ured PVCs is not changed, the network answers with a normal STATUS message;
it is used only to verify the link integrity (this message doesn’t contain the state of
the configured PVCs because nothing is changed);
– if one (or more) PVC has been added to or deleted from the FRL, or if the state
of any configured PVCs is changed, the network answers with a FULL STATUS
message, reporting the status of ALL the PVCs (this message is also used to
verify the link integrity);
3. after N391 polling cycles (i.e., after N391 expirations of the T391 timer), the PCU
sends to the network a STATUS_ENQUIRY message requiring a FULL STATUS
answer;
4. every time the PCU doesn’t receive an answer from the SGSN, an “error counter” is
incremented;

154 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

5. if the “error counter” reaches the N392 value during the error observation window
defined by:

N393 * T391

the Frame Relay link is put into Disable state, and all the contained PVCs are, as a
consequence, put in Disable state;
6. if the N392 threshold is not reached during the error observation window, the “error
counter” is restarted.

PCU SGSN

Expiration of STATUS_ENQUIRY
T391

STATUS
Reset and
restart T392

Expiration of STATUS_ENQUIRY
T391

STATUS
Reset and
Expiration of restart T392
T391
STATUS_ENQUIRY
N391 polling
cycles reached
FULL_STATUS

Fig. 7.15 Periodic Polling Procedure

The value of the T391 timer set on the BSC side must be lower than the value of the
i T392 timer set on the SGSN side.

7.2.2 Network Service Control


The Network Service Control entity is responsible for the following functions:
• NS SDU transmission: the NS SDUs are transmitted on the configured NSVCs.
The NS SDUs are encapsulated into Network Service Control PDUs which in turn
are encapsulated into Sub- Network Service PDUs. On each NSVC, data is trans-
ferred in order;
• Load sharing: the load sharing function distributes the NS SDU traffic among the
available (i.e., unblocked) NSVCs.
• NSVC management:
– a blocking procedure is used by a NS entity to inform an NS peer entity when an
NSVC becomes unavailable for NS user traffic;
– an unblocking procedure is used for the reverse operation;

A30808-X3247-L24-5-7618 155
GPRS/EGPRS Global Description Information
System

– a reset procedure is used between peer NS entities in order to set an NSVC to a


determined state, after events resulting in possibly inconsistent states of the
NSVC at both sides of the Gb interface;
– a test procedure is used to check that an NSVC is properly operating between
peer NS entities.

7.2.2.1 Load Sharing


All NS SDUs to be transmitted over the Gb interface are passed to the load sharing func-
tion. The load sharing function is used by NS entities to select, among unblocked
NSVCs of the addressed BVC, where to send NS SDUs.
Each BVC represents a GPRS/EGPRS cell in the PCU (see "7.3 BSSGP Protocol"); the
i load sharing function allows transmission of the NS SDUs related to a cell among the
available NSVCs.

The mapping between NS SDUs and NSVCs is based on an implementation dependent


function that meets the following requirements:
• for each BVC (i.e., for each cell), the load sharing function chooses the NSVC over
which the current NS SDU must be transmitted;
• thus, the load sharing function guarantees that, for each BVC, the order of all NS
SDUs is preserved;
• load sharing functions at the BSS and the SGSN are independent. Therefore, uplink
and downlink NS SDUs for a subscriber may be transferred over different NSVCs;
• a change in the number of available NSVCs for NS user traffic (i.e., one or more
NSVCs become blocked or unblocked) results in a reorganization of the NS SDU
traffic among the unblocked NSVCs;
• for a BVC, when there are no unblocked NSVCs between a BSS and a SGSN, the
corresponding traffic is discarded by the NS at the sending side.
Load sharing applies only to NS SDUs, not to NS signalling, such as NSVC manage-
i ment PDUs (e.g., NS_BLOCK_PDU used in the NSVC block/unblock procedures, see
"7.2.2.2 Control Procedures").

7.2.2.2 Control Procedures


The procedures concerning the management of NSVCs are:
• block/unblock of NSVCs by the operator
• reset and test the status of NSVCs.
The Block Procedure inhibits an NSVC from carrying traffic. For instance, the BSC may
block a NSVC because of:
– Operation and Maintenance intervention at the Gb interface, making the NSVC
unavailable for NS user traffic;
– equipment or link failure at the BSS or at the SGSN side;
– failure in the transit network.
The Load Sharing function is then informed and the result is the redistribution of NS SDU
to other unblocked NSVCs; the NS entity is informed via NS_STATUS_INDICATION
primitive for each affected BVC, while the remote peer is notified via NS_BLOCK_PDU.
The reception of the NS_BLOCK_ACK primitive from the SGSN closes the procedure
at the BSS side.
When the PCU has sent an NS_BLOCK_PDU, it waits TNSVCBLK seconds for
acknowledgement from the SGSN. The NNSVCBLKR parameter specifies the

156 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

maximum number of performed retries by the PCU in the NSVC block procedure; i.e., if
the SGSN does not answer the block procedure, the PCU retries the procedure at most
NNSVCBLKR times.
The Unblock Procedure allows the return of a previously blocked NSVC back to
service. The procedure is analogue to the BLOCK one.
When the PCU has sent the NS_UNBLOCK_PDU, it waits TNSVCBLK seconds for
acknowledgement from the SGSN. The NNSVCUBLR parameter specifies the
maximum number of performed retries in the NSVC unblock procedure; i.e., if the SGSN
does not respond to the unblock procedure, the procedure is retried NNSVCUBLR
times.
The Reset Procedure is used:
– when a new NSVC is set up between a BSS and the SGSN;
– after processor restart;
– after failure recovery or any local event restoring an existing NSVC which was in
dead state;
– when the state of an NSVC is undetermined between remote NS entities.
Upon completion of the reset procedure, the successfully reset NSVC is marked as
blocked and alive at both sides of the Gb interface.
The BSS (or the SGSN) sends the NS_RESET_PDU to its peer entity indicating the
NSVCI. The NS_RESET_PDU is sent on the NSVC being reset.
After the PCU sends the NS_RESET_PDU, it waits TNSVCR seconds for acknowledge-
ment. The NNSVCRR parameter specifies the maximum number of performed retries in
the NSVC reset procedure, before generating any alarm; i.e., if the SGSN does not
respond to the reset procedure, the procedure is retried infinitely times, but after
NNSVCRR times an O&M alarm is generated.
The Test Procedure is performed via NS_ALIVE_ACK_PDU and it is used when a BSS
(or SGSN) wishes to check that end-to-end communication with its peer entity exists on
an NSVC. Both sides of the Gb interface must initiate this procedure independently from
each other. This procedure is initiated upon successful completion of the reset proce-
dure (as specified in sub-clause "Reset procedure") and will then be periodically
repeated. After unsuccessful attempts, the procedure is stopped; the NSVC is marked
as dead and blocked and the O&M system and the load sharing function are informed.
A blocking procedure is initiated using an alive NSVC, if any.
The test procedure is executed according to the following features:
– the periodicity of the procedure is given by the TNSVCTST timer; i.e., when an
NSVC is available, the test message is sent to the SGSN every TNSVCTST
seconds;
– if after TNSVCPTST seconds no answer to the test is received from the SGSN, the
procedure is retried;
– after NNSVCTSTR repetitions, without any answer, the link is declared not available.

7.3 BSSGP Protocol


The Gb interface allows many users to be multiplexed over a common physical
resource. Both GPRS/EGPRS signalling and user data may be sent on the same phys-
ical resource.
The primary functions of the BSSGP protocol include:

A30808-X3247-L24-5-7618 157
GPRS/EGPRS Global Description Information
System

• transmit LLC frames from the SGSN to the BSS, with radio related information (such
as Quality of Service and routing information) which is used by the RLC/MAC func-
tion;
• transmit LLC frames from the BSS to the SGSN, with radio related information (such
as Quality of Service and routing information) which is derived from the RLC/MAC
function;
• provide functionalities to enable both the SGSN and the BSS to perform manage-
ment control functions (e.g., SGSN-BSS flow control).

7.3.1 BSSGP Addressing: BSSGP Virtual Connections (BVCs)


The BSSGP protocol establishes the connection between the SGSN and the PCU in
terms of BSSGP virtual connections (BVCs).
Each BVC is used in transporting BSSGP PDUs between PTP (point to point) functional
entities; a PTP functional entity is constituted by a cell. Note following concepts:
• each BVC is identified by a BVCI (Virtual Connection Identifier);
• each BVCI identifies univocally a GPRS/EGPRS cell in the PCU;
• each PCU is identified by the Network Service Entity Identifier (NSEI) in the SGSN;
• the couple BVCI and NSEI identifies univocally a GPRS/EGPRS cell in the SGSN;
In GSM 08.18 – for point-to-point packet transfer – it’s specified that a cell is identified
by a BVCI, so there is a one to one relationship between a cell and a BVCI.
In Siemens implementation, the PTPPKF object represents the presence of packet
switched data services in a specific cell and the state of this object allows or disallows
the service in the cell.
The dependency from the BTS object is one to one; then all state changes on BTS
objects are reflected on PTPPKF objects.
The NSVC-BVCI hierarchy is not one to one but one PTPPKF can be reached from
different NSVCs, which are connected to the same PCU; in fact, the NSEI identifies the
NSVC group, i.e., the group of all the NSVCs that provide service for a PCU: to one PCU
corresponds only one NSVC group and vice versa.
The PTPPKF state can be affected by:
– BTS state changes;
– specific commands executed on the object;
– state changes on subordinated NSVC objects.
The PTPPKF state transition, due to NSVC state change, is handled by the system that
puts the PTPPKF object into the disabled state when the associated BVCI is no longer
accessible. All state transitions for PTPPKF objects are notified to the remote end via
the BVCI Block/Unblock procedure (see 7.3.1.1). The PCU restart or the BSS initializa-
tion are handled generating a reset procedure with the SGSN. Needed timers for
handling Block/Unblock and Reset procedures are defined in the PCU object (see
"7.3.1.1 BVC Procedures").
The PTPPKF object is always created with the instance equal to 0 since it is subordinate
to the BTSM and BTS object following this path:
BTSM:m/BTS:n/PTPPKF:0
The BVCI number associated to a PTPPKF object instance is fixed, and the relation is:

158 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

BVCI number = (number of creation of the PTPPKF in the database) + 2

The 0 and 1 values are reserved respectively for signalling and PTM links.
When an upgrade from the Release BR5.5 to BR7.0 is executed, some changes in the
i SGSN database must be executed. This is due to the fact that, according to the load
balancing schema that is used for the PCUs (see "6.1 Supported BSC Types"), the
PTPPKFs (i.e., the BVCIs) are no longer statically assigned to a single NSEI (i.e., to a
single PCU) but they can be moved from one PCU to another one following the PTPPKF
distribution algorithm (see "8 Load Control for Packet Switched Services"); so in the
SGSN, the BVCIs of one BSC have to be configured on all the NSEIs (PCUs) related to
the BSC.

To summarize the previous concepts, let us consider a SGSN that manages four PCUs:
• PCU:0, PCU:1 and PCU:2 configured on BSC:1;
• PCU:0 configured on BSC:2.
Reflected in Fig. 7.16, it can be seen that:
a) each PCU is identified in the SGSN by the NSEI attribute:
– the PCU:0 of the BSC:1 is identified by the NSEI_A value;
– the PCU:1 of the BSC:1 is identified by the NSEI_B value;
– the PCU:2 of the BSC:1 is identified by the NSEI_C value;
– the PCU:0 of the BSC:2 is identified by the NSEI_D value;
b) the NSEI attribute also identifies all the configured NSVCs for each PCU:
– the NSEI_A value identifies NSVC:0, NSVC:1 and NSVC:2 connections (related
to PCU:0 of BSC:1);
– the NSEI_B value identifies NSVC:3 and NSVC:4 connections (related to PCU:1
of BSC:1);
– the NSEI_C value identifies NSVC:5, NSVC:6 and NSVC:7 connections (related
to PCU:2 of BSC:1);
– the NSEI_D value identifies NSVC:0, NSVC:1 and NSVC:2 connections (related
to PCU:0 of BSC:2);
Obviously, the NSVCI values, related to the different NSVCs created for
i the four PCUs must be different from each other.

c) each cell is identified in the PCU by the BVCI value;


d) each cell is identified in the SGSN by the couple BVCI and NSEI;
e) the traffic of all the cells (BVCIs) configured for a PCU is distributed among all the
NSVCs configured for the PCU.

A30808-X3247-L24-5-7618 159
GPRS/EGPRS Global Description Information
System

NSVC group identified by NSEI_A


GPRS
Cell
FRL:0
BVCI= 2
NSVC:0
BVCI=2
GPRS BSC:1\PCU:0 NSVC:1
Cell NSEI_A
NSEI_A BVCI=6
BVCI= 6 NSVC:2

FRL:1

NSVC group identified by NSEI_B


GPRS FRL:2
Cell
BVCI= 5 BSC:1\PCU:1 NSVC:3 BVCI=4
NSEI_B NSVC:4 NSEI_B
BVCI=5
GPRS
Cell
BVCI= 4 FRL:3

NSVC group identified by NSEI_C SGSN


GPRS FRL:4
Cell
BVCI= 3 NSVC:5
BVCI=3
BSC:1\PCU:2 NSVC:6 NSEI_C
NSEI_C BVCI=7
GPRS NSVC:7
Cell
BVCI= 7
FRL:5

GPRS NSVC group identified by NSEI_D


Cell FRL:0
BVCI= 3
NSVC:0 BVCI=2
GPRS BSC:2\PCU:0 NSVC:1
Cell BVCI=3 NSEI_D
NSEI_D
BVCI= 2 NSVC:2
BVCI=4

GPRS FRL:1
Cell
BVCI= 4

Fig. 7.16 Distribution of Packet Switched Data Traffic among Different Cells

160 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

7.3.1.1 BVC Procedures


For BVCs associated to PTPPKF objects, the following procedures with remote end are
implemented:
• BLOCK;
• UNBLOCK;
• RESET.
The BLOCK Procedure inhibits a BVCI from carrying traffic. It’s performed when the
PTPPKF object is locked by the operator or when it reaches a disable-dependency
state. All PDTCHs of the cell are released and system information, reporting
GPRS/EGPRS service not allowed in the cell, is sent in the BCCH or the PBCCH.
If, after a Block Procedure attempt, the PCU doesn’t receive a response from the SGSN,
it retries the procedure. The waiting time for the block procedure is defined by the T1
parameter: after sending a BVCI block message, the PCU waits T1 seconds for
acknowledgement. After NBVCBR consecutive repetitions, without any answer from the
SGSN, an O&M alarm is sent.
The UNBLOCK Procedure allows the traffic back on the previous blocked BVCI. The
PTPPKF is put in enabled state. System information, reporting GPRS/EGPRS service
allowed in the cell, is sent in the BCCH or the PBCCH.
If, after an Unblock Procedure attempt, the PCU doesn’t receive a response from the
SGSN, it retries the procedure. The waiting time for the unblock procedure is defined by
the TF1 parameter: after sending a BVCI unblock message, the PCU waits T1 seconds
for acknowledgement. After NBVCUR consecutive repetitions, without any answer from
the SGSN, an O&M alarm is sent.
The RESET Procedure is used when a new BVCI is set up between the SGSN and the
BSS, or after each event (processor restart, failure recovery, etc.) that needs to clear
and to synchronize BVCI status on both sides. Both sides must initiate the procedure
independently.
If, after a Reset Procedure attempt, the PCU doesn’t receive a response from the SGSN,
it retries the procedure. The waiting time for the reset procedure is defined by the T2
parameter: after sending a BVCI reset message, the PCU waits T2 seconds for
acknowledgement. After NBVCRR consecutive repetitions, without any answer from the
SGSN, an O&M alarm is sent.

7.3.2 Quality of Service (QoS)


For an uplink data transfer, the QoS profile is communicated by the MS as a priority
information in the PACKET_CHANNEL_REQUEST message. For a downlink data
transfer, the BSSGP protocol provides the means to transfer the full QoS profile together
with each downlink LLC PDU. In the latter case, the following QoS parameters are
included in each LLC-PDU transferred to the BSS:
– precedence class;
– peak throughput;
– LLC-PDU lifetime.
The PCU, taking into account the available radio resources and the multislot capabilities
of the MS, decides if and how the requested QoS may be satisfied. This means that the
core algorithm of the PCU would try to satisfy the requested QoS by acting on many
factors.

A30808-X3247-L24-5-7618 161
GPRS/EGPRS Global Description Information
System

Regarding the QoS, as described in "5.3 Management of Packet Data Channels", the
resource allocation algorithm allows the consideration of the required peak throughput
class.
No QoS related to BSSGP flow control ("7.3.3 SGSN-BSS Flow Control") is now imple-
mented.

7.3.3 SGSN-BSS Flow Control


Packet switched data traffic exhibits large statistical fluctuations both for the flow-in into
the PCU as well as for the flow-out over the air interface. The former depends on the
volatility of Internet traffic; the latter is caused by the speciality of the air interface.
Strongly varying C/I leads to variations of the RLC blocks re-transmission rates. GSM
voice calls have pre-emptive priority and may thus steel time slots allocated to PS
services (but not the GPRS/EGPRS reserved timeslots).
For these reasons two different types of flow control are implemented in the BSS:
– BVC Flow Control: the variance both of the inter-arrival times of arriving frames and
deleted RLC blocks causes waiting time in the BVC buffer; as a consequence rate
control schemes are used to smooth traffic and thus reduce variance. This reduces
transmission delays for the user. Additionally, flow control should minimise the prob-
ability of BVC buffer over flow.
– MS Flow Control: without mobile specific flow control, mobiles with low individual
flow rate (caused for example by being capable of only one timeslot, multiplexed
with other TBFs onto the same timeslot, and large retransmission rates) might slow
down all other mobiles within the same cell. Because these mobiles cannot get their
data out of the common BVC buffer fast enough, the buffer filling might increase
above a threshold, and the mean BVC flow rate R might slow down. Thus, the SGSN
might narrow the throttle for all other mobiles, too.
For packet transfer from the SGSN to the BSS, the BSSGP protocol uses an address
which is composed of three parts:
• cell Identity (BVCI);
• QoS profile;
• MS identification (e.g. TLLI).
The Temporary Logical Link Identity (TLLI) identifies univocally a GPRS/EGPRS user,
i who is engaged in data transfer, inside a cell (see also "9.8.2.5 Contention Resolution").

These three parts are then used to dynamically queues and contexts in both the SGSN
and the BSS. The flow control mechanism is then based on these queues and contexts.
The principle of flow control is based on the following:
a) in the SGSN, queues are provided per MS. The SGSN sends PDUs to the LLC layer
as a function of the requested service type and the Mobility Management state (see
"9.3.1 Mobility Management States");
b) in the BSS, queues per cell (BVC) and per MS (TLLI) are provided at the BSSGP
level;
c) signalling has its own queue.
The BSS controls the flow of packet data units (PDUs) to its BVC buffer for an individual
MS, by indicating to the SGSN the maximum allowed throughput for a certain TLLI.
The BSS controls the flow of packet data units to its BVC buffers by indicating to the
SGSN the maximum allowed throughput for each BVC.

162 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

The amount of buffered packet data units for a given TLLI or BVC has to be optimized
to efficiently use the available radio resources. The packet data units have to be trans-
ferred across the Um interface before the PDU lifetime expires; in this case, the PDU is
deleted from the BSS and the deletion is signalled to the SGSN by the LLC-
DISCARDED PDU message.
It is foreseen a cascaded mobile (MS) and cell (BVC) oriented “flow control scheme”
(see Fig. 7.17) for the downlink transmission of LLC frames from the SGSN to the PCU
(in uplink transmission the problem does not exist since it is the BSS itself which sched-
ules the MS accesses, according to its own radio capacity). A LLC PDU must first have
obtained the permission of the mobile flow control before it is submitted the cell (BVC)
specific flow control.

Fig. 7.17 Cascaded Flow Control

The Token Bucket Algorithm used in the Flow Control procedure works in the following
way (see Fig. 7.18): there is a queue of LLC frames without a permit for transmission to
the PCU, and a bucket of permits (“tokens”). The LLC frame at the head of the frame
queue obtains a permit if at least one token is available in the permit bucket. In this case,
it joins the buffer of LLC frames with permits waiting to be transmitted, and the token is
deleted. Permits are generated at the rate R as long as the number in the permit bucket
does not exceed a certain threshold Bmax. When frames have different sizes, a token
should be thought as the permission to transmit one byte. A frame p of size L(p) will
obtain the permit for transmission, if at least L(p) tokens are available in the bucket.

A30808-X3247-L24-5-7618 163
GPRS/EGPRS Global Description Information
System

Fig. 7.18 Token Leaky Bucket (in SGSN)

On the PCU side, there is for each BVC or MS a buffer which is filled by the segmenta-
tion of the arriving LLC PDUs and empties when these blocks are transmitted over the
air interface. PCU calculates the control variables R and Bmax and transmits them with
flow control commands to the SGSN at every expiration of the TF1 timer. Thus a closed-
loop control is realised (see Fig. 7.19). In the PCU the real rate Rpcu (towards the Abis
interface) can be different from the value R sent to SGSN.
In other words, SGSN uses parameters sent by BSC in order to decide if it can send
data or not. The principle is that periocally BSC can send new parameters and SGSN
updates the internal values related.

164 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Fig. 7.19 Closed Loop Control

Every TF1 timer expiration, the BSC can send a new BVC or MS Flow Control with
following parameters updated:
– Bucket Size (Bmax);
– Bucket Leak Rate (R);
– Bucket FullRatio.
PCU and SGSN are provided with two different types of buffer, one for the BVC Flow
Control and another one for the MS Flow Control. As a consequence, the following
parameters are defined:
– BmaxPCUBVC is the maximum size of the buffer in PCU for the BVC Flow Control;
– BmaxPCUMS, is the maximum size of the buffer in PCU for the MS Flow Control;
– BmaxBVC is the maximum size of the buffer in SGSN for the BVC Flow Control;
– BmaxMS is the maximum size of the buffer in SGSN for the MS Flow Control.
The Bucket Leak Rate (R) is the rate at which the permits for transmission are gener-
ated.
The Bucket full ratio represents the percentage of the Current Bucket Level (Buck-
etLevelPcu) compared to Bmax.
The BSS has to trigger the Flow Control message in a way that the BSS can guarantee
a continuous data flow to the MSs.
Following things have to be taken into account:
– a too low Leak rate respect to the maximum rate possible prevents SGSN from
sending data to BSC;
– Bmax has to be high enough in order to guarantee that BSC has always “something
to send”;
– Bmax has to be low enough due to PDU-lifetime, for this reason it is better to have
a little Bucket in order to have a minimum permanence in the BSC Bucket;

A30808-X3247-L24-5-7618 165
GPRS/EGPRS Global Description Information
System

– the discarding of packets because of lifetime expiration will be an exceptional case.


The current Bucket Level procedure is used only if SISGSNREL99 parameter (PCU
i object) is set to the TRUE value.

7.3.3.1 MS Flow Control Message


During active Downlink TBF, the BSC sends a MS-FLOW-CONTROL (see Fig. 7.19)
message in order to inform SGSN about the quantity of memory reserved for that TBF
and the rate at which bytes are sent.
At each TF1 timer expiration, the BSC computes BmaxBasicMS value according to the
following formula:

BmaxBasicMS = f *(1s +TF1) * RmaxMS

where:

f=1 if no large amount of LLC lifetime expiration occurs;


f =1/ ((nGPRS_TS +1) * TF1 * i) otherwise.

Besides:
• the value i represents the consecutive number of times that the LLC PDU lifetime
expiration threshold is found to be present. LLC PDU lifetime expiration threshold is
reached when in the previous interval the number of bytes related to expired LLC is
more than 30% of LLC bytes sent for that MS;
• nGPRS_TS is the number of timeslots assigned for GPRS/EGPRS services at timer
expiration at this MS;
• RmaxMS is the teorical MS Maximum Rate according to resources assigned to MS
and to used coding scheme (CS/MCS); it is defined as:

K
RmaxMS = ∑ TSPercentageR TSk

k=1

where:
- K represents the Number of timeslots assigned to the MS;
- RTSK is the rate in case the entire timeslot is assigned to that MS;
- TSPercentage is a percentage that indicates how the timeslot is exploited (in
percentage) by the MS, when it shares the timeslot with other MSs. For example
TSPercentage=30% means that 30/100 * RTS is the rate for the MS in that timeslot.

In order to allow more flexibility in flow control management, some parameters are intro-
duced.
The purpose of these parameter is for testing purposes and for special application. It will
i not be guaranteed that the BSC works with all combination and it is suggested to
mantain default value.

One of these parameters is MSBMAPER (MsBucketMaxPercentage), defined as:

166 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

BmaxMS = MSBMAPER * BmaxBasicMS


Through this parameter the operator can reduce the value of BmaxBasicMS automati-
cally computed by BSC.
Another parameter is called MSBSPPER (MsBucketSizePcuPercentage) and repre-
sents a relationship between BmaxPCUMS and BmaxMS . The exact definition is:

BmaxPCUMS = MSBSPPER *BmaxMS

As can be seen in Fig. 7.19, BmaxPCUMS must be higher than BmaxMS, so MSBSPPER
is greater than 100.
The states “congested” and “non-congested” are determined using two thresholds:
BhighMS and BlowMS. Starting from BmaxPCUMS the operator can define these two
thresholds through the parameters: MSBHIPER (MsBucketHighPercentage) and
MSBLPER (MsBucketLowPercentage). They are defined through the following formula:

BhighMS = MSBHIPER * BmaxPCUMS


BlowMS = MSBLPER *BmaxPCUMS

If the state is “non-congested”, and BucketLevelPCUMS (the current Bucket Level in


PCU) crosses BhighMS from below, the state is set to “congested”.
If the state is “congested”, and BucketLevelPCUMS crosses BlowMS from above, the
state is set to “non-congested”.
For what concerns the MS Bucket Leak Rate, different behaviours are foreseen,
according to these two situations:
– Bucket_Full Ratio is not implemented in SGSN;
– Bucket_Full Ratio is implemented in SGSN.
The value of Bucket Full Ratio is the percentage of BucketLevelPCU compared to
Bmax. For example if BucketLevelPCU is 40% of Bmax, Bucket Full Ratio = 40. If
BucketLevelPCU is 150% of Bmax, Bucket Full Ratio = 150.

Bucket_Full Ratio is not implemented in SGSN


The Flow Rate out of the buffer in the SGSN (RMS) can be expressed in the following
way:

MS Bucket Leak rate = RMS*8/100

(8 is due to the fact that Bucket Leak rate is in 100 bits/s unit, Bucket size is 100 octet
increments).
In the normal case the BSC sends to SGSN the following value for RMS:

RMS= RmaxMS

Otherwise the BSC sends the following value:

A30808-X3247-L24-5-7618 167
GPRS/EGPRS Global Description Information
System

RMS=f* RmaxMS

where the coefficient f is chosen according to the following situations:


1. if Leak Rate has not been already reduced for Bucket Congestion, then the leak rate
is reduced as following:

Retrasm_rate = Bytes retransmitted/ Byte transmitted

if Retrasm_rate < 10 % no correction.


if 10 % < Retransm_rate < 20 % then Leak rate = 0.9 * Leak Rate
if 20 % < Retransm_rate < 30 % then Leak rate = 0.8 * Leak Rate
if 30 % < Retransm_rate < 40 % then Leak rate = 0.7 * Leak Rate
if 40 % < Retransm_rate < 50 % then Leak rate = 0.6 * Leak Rate
if 50 % < Retransm_rate < 60 % then Leak rate = 0.5 * Leak Rate
if 60 % < Retransm_rate < 70 % then Leak rate = 0.4 * Leak Rate
if 70 % < Retransm_rate < 80 % then Leak rate = 0.3 * Leak Rate
if 80 % < Retransm_rate < 90 % then Leak rate = 0.2 * Leak Rate
if 90 % < Retransm_rate < 100 % then Leak rate = 0.1 * Leak Rate
2. If MS buffer is congested:

f = 1/((nGPRS_TS +1) * TF1 * i)


where:
– i is the consecutive number of times that the LLC PDU lifetime expiration
threshold is found to be present;
– nGPRS_TS is the number of TS assigned for GPRS/EGPRSD services at timer
expiration for this MS.

Bucket_Full Ratio is implemented in SGSN


In this case, the value MS Bucket Leak rate, is sent always without correction factor.

MS Bucket Leak Rate = RmaxMS

7.3.3.2 BVC Flow Control Message


BmaxBVC (see Fig. 7.19 ) is a value not corresponding to real BVC memory capacity on
PCU, but a value to be communicate to SGSN that it can send downlink LLCs.
The same criteria and parameters will be applied as in BmaxMS case. More specifically,
BSC calculates the BmaxBasicBVC according to the following formula:

BmaxBasicBVC= f*(1s + TF1) * RmaxBVC

where:

168 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

f=1 if no large amount of LLC lifetime expiration occurs;


f =1/ ((nGPRS_TS +1) * TF1 * i) otherwise.

Besides:
• the value i represents the consecutive number of times that the LLC PDU lifetime
expiration threshold is found to be present. LLC PDU lifetime expiration threshold is
reached when in the previous interval the number of bytes related to expired LLC is
more than 10% of LLC bytes sent for all BVC;
• nGPRS_TS is the number of timeslots assigned for GPRS/EGPRS services at timer
expiration;
• RmaxBVC depends on the number of timeslots that can be assigned to
GPRS/EGPRS, and on the data rate on these timeslots, basing on initial CS or initial
MCS (see 10.5.3) or in case Link Adaptation is enable on the maximum CS or MCS;
it is defined as:

K N PCCH KEDGE KGPRS


Rmax B V C = ∑R TSk = ∑ RTSk + ∑ RTSk + ∑ RTSk
k=1 k=1 k=1 k=1

where:
K = NPCCH + KEDGE + KGPRS
NPCCH: is the number of configured packet control channels;
KEDGE: maximum number of channels (configured, unlocked and
enabled) can be assigned to EDGE after applying GPDPDTCHA
parameter.
KGPRS: maximum number of channels (configured, unlocked and
enabled) can be assigned to GPRS (but not to EGPRS) after
applying GPDPDTCHA parameter. In this number timeslots
belonging to EDGE capable TRXs are not considered (they are
counted in the KEDGE value).

Besides, the following definitions are necessary:

Ndinamic: maximum number of channels (configured, unlocked and enabled)


can be theoretically used to either PS services or CS services in the
cell;
NdinamicEDGE: maximum number of channels (configured, unlocked and enabled)
can be theoretically used for EGPRS service on EDGE capable
TRXs.
The number refers to timeslots without applying GPDPDTCHA
parameter.
NdinamicGPRS: maximum number of channels (configured, unlocked and enabled)
can be theoretically used for GPRS service on GPRS capable
TRXs.
The number refers to timeslots without applying GPDPDTCHA
parameter.

A30808-X3247-L24-5-7618 169
GPRS/EGPRS Global Description Information
System

Nreserved : number of PDCH reserved for GPRS/EDGE services. It corre-


sponds to GMANPRES parameter.
GPDPDTCHA : percentage of channel available can be assigned to GPRS or EDGE
(if possible).
K_EG: maximum number of channels (configured, unlocked and enabled)
can be dynamically assigned to GPRS/EGPRS services. The
number refers to timeslots after applying GPDPDTCHA parameter.
To understand previous definitions let us suppose that a cell (BVC) has the configuration
shown in Fig. 7.20.

TDMA frame
GPRS
TRX 0 BCCH SDCCH PBCCH
Capable
0 7
TDMA frame
EGPRS
TRX 1 SDCCH
Capable
0 7
TDMA frame
EGPRS
TRX 2 SDCCH
Capable
0 7

GMAPERTCHRES=4
GPDPDTCHA=50%
Configured Packet Control Channels=1

Fig. 7.20 Example Cell Configuration

In this case we have the following values:


– NPCCH = 1
– Ndinamic =19
– NdinamicEDGE =7
– NdinamicGPRS =12
For the RmaxBVC computation, the following information must be taken into consider-
ation:
– If Ndinamic is greater than Nreserved, the parameter K_EG is equal to GPDP-
DTCHA*(Ndinamic - Nreserved), otherwise if Ndinamic is lower or equal to
Nreserved, then K_EG is equal to zero.
– K = NPCCH + K_EG + Nreserved
– If K_EG + Nreserved is greater than NdinamicEDGE, then:
KGPRS = K_EG + Nreserved - NdinamicEDGE
KEDGE = K_EG + Nreserved - KGPRS
otherwise
KGPRS = 0
KEDGE = K_EG + Nreserved

Example 1

170 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Configuration as in Fig. 7.20, if all channels are enabled/unlocked (GPDP-


DTCHA=50%).

NPCCH = 1
Ndinamic =19
NdinamicEDGE =7
NdinamicGPRS =12
Nreserved =4

K_EG = 0.5 *(19-4)=0.5*(15)=8


KGPRS = K_EG + Nreserved - NdinamicEDGE = 8 + 4 - 7 = 5
KEDGE = K_EG + Nreserved - KGPRS = 8 + 4 - 5 = 7

K = 1 + 8 + 4 = 13

Example 2
The TRX1 becomes disabled (GPDPDTCHA=50%).

NPCCH = 1
Ndinamic =12
NdinamicEDGE =7
NdinamicGPRS = 5
Nreserved =4

K_EG = 0.5 *(12-4)=0.5*(8)=4


KGPRS = K_EG + Nreserved - NdinamicEDGE = 4 + 4 - 7 = 1
KEDGE = K_EG + Nreserved - KGPRS = 4 + 4 - 1 = 7

K=1+4+4=9

Also in case of BVC flow control, two new parameters have been introduced BVCB-
MAPER (BvcBucketMaxPercentage) and BVCBSPPER (BvcBucketSizePcuPer-
centage). Through them the operator can modify the values automatically computed by
the BSC. The new parameters are defined in the following way:

BmaxBVC = BVCBMAPER * [(1+TF1) * RmaxBVC ]


BmaxPCUBVC = BVCBSPPER *BmaxBVC.

The states “congested” and “non-congested” are determined using the two thresholds
BhighBVC and BlowBVC of the BmaxPCUBVC.
Two parameters allow the operator to express them as a function of BmaxPCUBVC:

BVCBHIPER (BvcBucketHighPercentage) and BVCBLPER (BvcBucketLowPer-


centage). These new parameters are contained in the following formulas:

A30808-X3247-L24-5-7618 171
GPRS/EGPRS Global Description Information
System

BhighBVC = BVCBHIPER * BmaxPCUBVC


BlowBVC = BVCBLPER * BmaxPCUBVC

If the state is “non-congested”, and BucketLevelPCUBVC (the current bucket level in


PCU) crosses BhighBVC from below, the state is set to “congested”.
If the state is “congested”, and BucketLevelPCUBVC crosses BlowBVC from above, the
state is set to “non-congested”.
For what concerns the BVC Bucket Full Ratio, the same criteria as in section 7.3.3.1 are
applied.

7.3.3.3 Flow Control sending criteria (for both BVC and MS)
A BVC Flow Control/MS Flow Control can be sent at each TF1 timer expiration. In order
to reduce the number of FLOW-CONTROL messages sent, they will be sent only when:
– Bmax or R is changed compared to the previous parameter sent, i.e., in case of too
many PDU lifetime expiration or resource increased/decreased;
– If Bucket Ratio is not implemented, in case of too many RLC retransmission or
BucketLevelPCU exceeds Bhigh or goes below Blow threshold;
– If Bucket Ratio is implemented:
- every time that BucketLevelPCU is more than 70% of Bmax (this means conges-
tion at BSS side);
- every time that BucketLevelPCU is less than 5% of Bmax. This behaviour is
assumed in order to prevent possible misalignment between Bucket at SGSN side
and at BSC side; it is possible that bucket in BSC is more or less zero and completely
full at SGSN side due to SW error whatever.
If SGSN does not answer to BVC FLOW control, the PTPPKF object is put disable and
BVC RESET procedure starts.
If SGSN does not answer to MS Flow Control, BSC stops sending FLOW_CONTROL
messages for that TBF.
Fig. 7.21 and Fig. 7.22 show the message flow related to MS-Control Flow for two
different cases: the normal case and the case in which the SGSN does not answer.

172 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Fig. 7.21 MS-FLOW-CONTROL

Fig. 7.22 SGSN does not answer with MS-FLOW-CONTROL-ACK message

A30808-X3247-L24-5-7618 173
GPRS/EGPRS Global Description Information
System

It must be noted that in case redundant NSVC links are created on the Gb-interface, the
following rule must be obeyed to avoid unnecessary GPRS/EGPRS blocking for certain
cells (see "7.2.1.3 Procedures for PVCs"): if NS-ALIVE-ACK is not received because of
link problems, the respective NSVC is put into disabled state after a maximum time
defined by:

Time_1= TNSVCTST+ TNSVCPTST * NNSVCTSTR

The PTPPKF object is put into disabled state if FLOW-CONTROL-ACK is not received
during the time defined by:

Time_2 = TF1* Number of Flow Control Retries

In case of link problems it could therefore happen that the PTPPKF object is disabled
while the NSVC is still enabled. To avoid this situation, the following rule has to be
followed:

Time_2 > Time_1

The default value of "Number of Flow Control Retries" is fixed to the value 15. The rule
is therefore fulfilled with the database default values for the following parameters:
– TNSVCTST;
– TNSVCPTST;
– NNSVCTSTR;
– TF1.

174 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

8 Load Control for Packet Switched Services


As described in "6.1.4 PPCU and PPXU Redundancy and Configuration Rules", when
the user configures a GPRS/EGPRS cell, i.e., when the user creates a PTPPKF object
instance, he does not have to assign the cell to a specific PCU, but it is the system that
dynamically assigns the cell to one of the PCUs. This is a direct consequence of the load
balancing redundancy.
When one PCU fails (e.g., when a PPXU fails in cases of high capacity BSCs, or when
a couple of PPCUs fail in cases of standard BSCs), the system redistributes the PS
traffic among the remaining PCUs.
In the BSC, there is a load control system that manages the PTPPKFs distribution
among the PCUs available in the system.
This kind of calculation is established taking into account the number of created and
available PCUs, and the number of PTPPKF already associated to each PCU.
In case one PCU becomes unavailable for any reason, all PTPPKFs belonging to that
PCU are moved to other PCUs; the algorithm evaluates new PTPPKFs distribution
considering the number of PTPPKF already present on each PCU and the number of
resources reserved to packet switched services for each PTPPKF (see "8.1 Dynamic
PTPPKF Reconfiguration").
If one PCU becomes available the first time (i.e., after its creation), a new computation
is started in order to distribute PTPPKFs configured in the system on all PCUs.
Besides, the PTPPKF distribution process implements a mechanism based on a Guard
timer, in order to avoid oscillation in case of fast PCU state changes (see "8.1.5 Time
Needed to Execute PTPPKF Reconfiguration").
In order to enable the movement of one PTPPKF from one PCU to another, all PCUs
i must contain the same database as far as the PTPPKF configuration is concerned.

8.1 Dynamic PTPPKF Reconfiguration


The PTPPKFs distribution/redistribution algorithm is performed by the GPRS/EGPRS
load control process implemented on TDPC board.
The PTPPKFs reconfiguration is performed in the following cases:
• when the boards related to the PCU become unavailable; this situation could happen
when:
– a PPXU board fails, in cases of high capacity BSCs;
– a couple of PPCU boards fail (the couple PPCU:0/1 or the PPCU2/3), in cases of
standard BSC;
– the PPXU or the couple of PPCUs is locked.
• the connection of the PCU towards the SGSN goes down, that means that the last
NSVC connection from the PCU to the SGSN is no longer available; this situation
could happen when:
– the PCMG line containing the last available FRL (if the last available FRL is
created on it) is locked or goes down (as a consequence, the last NSVC becomes
unavailable);
– the PCMA line containing the last available FRL (if the last available FRL is
created on it) is locked or goes down (as a consequence, the last NSVC becomes
unavailable);

A30808-X3247-L24-5-7618 175
GPRS/EGPRS Global Description Information
System

– the last FRL is locked or goes down, and as a consequence, the last NSVC is
disabled;
– the last NSVC is locked or goes down;
– the PCU is locked;
• the connection of the PCU towards the SGSN comes back, that means that the last
NSVC connection from the PCU to the SGSN is now available.
In general, the PTPPKFs reconfiguration is triggered from all the operations
i that generate a PCU/Gb down-restart.
So the previous causes can be summarized with these sentences: “any time
the Gb interface related to any PCU is no longer available, the reallocation is
triggered” or “when a PCU <<can not see>> the SGSN, the PTPPKF allocated
to that PCU should be moved to another PCU that can <<see>> the SGSN“.

• a PTPPKF is deleted, but only if this operation causes an unbalanced distribution of


PTPPKFs among the PCUs.
It must be noted that:
1. the PTPPKF creation is not included as a trigger for PTPPKFs reconfiguration (when
a PTPPKF is created, it is assigned to the less loaded PCU);
2. the PCU creation itself does not trigger the PTPPKFs reconfiguration, but the recon-
figuration starts when the first NSVC is created and enabled for this PCU (i.e. when
the connection on the Gb interface is available);
3. the PCU overload (see "8.2 PCU Overload Management") does not trigger the
PTPPKF reconfiguration, since, in this case, no PCU/Gb down-restart is executed.
To handle the distribution/redistribution algorithm, the system uses the following infor-
mation:
• dynamic information:
– state of PCUs configured in the BSC;
– state of PTPPKFs configured in the BSC;
• static information:
– number of PCUs and PTPPKFs configured in the BSC;
– total number of radio channels configured for PS services for each PTPPKF
(packet control channels + reserved packet channels + dynamic packet chan-
nels);
– Routing Area of the PTPPKF (only at init time);
– total number of timeslots configured for each PCU on the Gb interface (see "7 Gb
Interface").
If we define:

PCU_TS_Gb Number of timeslots of 64 kbit/s related to the FR links associ-


ated to a specific PCU
PTPPKF_load(i) Number of GPRS/EGPRS channels (static and dynamic)
configured for a specific PTPPKF(i)

Then the load of a single PCU can be considered as the sum of the PTPPKF_load of
all the PTPPKF associated to the PCU divided by the PCU_TS_Gb of the PCU (that is
how the FR links associated to the PCU can manage all the radio channels associated
to the PCU); so:

176 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

PCU_load = (Sum PTPPKF_load(i) [i=1..n]) / PCU_TS_Gb


where n is the number of PTPPKFs associated to the PCU

Then the algorithm tries to distribute the packet switched traffic among the configured
and available PCUs, so that all the PCUs have the same PCU_load.
Moving one PTPPKF from one PCU to another one causes a release of all the TBFs
associated with that PTPPKF. To avoid, as much as possible, this situation, the set of
PTPPKF is divided in 3 sub sets (the three sets are considered by the general algorithm
that calculates and moves PTPPKFs from one PCU to another one):
• PTPPKF_DIED: this set contains PTPPKFs belonging to PCUs without the Gb in
service that have to be moved to PCUs in service; this set is taken into account as
soon as the algorithm runs after the PCU/Gb fault;
• PTPPKF_KO: it includes PTPPKFs that are not carrying traffic because they are
disabled or have been locked; this set is first analyzed for a possible moving of
PTPPKFs, when a new PCU goes into service.
• PTPPKF_ENABLE: it includes all the other PTPPKFs, that are considered for
possible moving; this set is the second analyzed for possible moving of PTPPKFs,
when a new PCU goes into service. The set is considered after the previous set (i.e.
PTPPKF_KO) has become void.
According to different situations, different handling is provided even if the general rule is
always to redistribute the traffic in the better way.
The following examples are shown:
– System initialization (Bring-Up and Full Init), see 8.1.1;
– Creation of a new PCU object (new PPXU board, or new PPCU couple of boards),
see 8.1.2;
– PCU crash (see 8.1.3);
– PCU comes back in service (see 8.1.4).

8.1.1 System Initialization


When the system is initialized, BVCI synchronization is performed, between the BSC
and the SGSN, using the BVC RESET procedure (see "7.3.1.1 BVC Procedures"). Then
the distribution algorithm (see Fig. 8.1) distributes the configured PTPPKFs among the
configured and available PCUs; the algorithm takes into account, besides the PCU
loads, the Routing Area of the PTPPKFs (regarding routing areas, see "9.2 Network
Structure").
The algorithm considers the Routing Areas only at init time.
i
In fact, to reduce the paging load, (if the SGSN is intuitive enough) the algorithm tries, if
possible, to associate a Routing area to just one PCU.
These constraints could collide with the balancing of the PCU loads. In this case, the
constraint related to the loads of the PCUs has an higher priority.

A30808-X3247-L24-5-7618 177
GPRS/EGPRS Global Description Information
System

PCU:0
PTPPKF
NSEI:0
RA=2 PTPPKF
PCU_load=9/3=3
1 PDCH RA=1
BVC=4 3 PDCH
BVC=23 PCU_TS_Gb= 3
PTPPKF
RA=2 PTPPKF PTPPKF
1 PDCH RA=2 RA=1
BVC=24 1 PDCH 3 PDCH
BVC=33 BVC=34

SGSN
PCU:4
NSEI:4
PCU_load=3/1=3 PTPPKF
RA=3
1 PDCH
BVC=77 PCU_TS_Gb= 1
PTPPKF
RA=5
2 PDCH
BVC=35

Fig. 8.1 Example of PTPPKF Distribution During System Initialization

8.1.2 Creation of a PCU Object and Enabling a NSVC for It


When a new PCU is created and the first NSVC is created and enabled, the algorithm
redistributes the already configured PTPPKFs among the available PCUs, taking into
account the new one.
On the Gb interface, the BVC BLOCK procedure (see 7.3.1.1) is performed on the
PTPPKFs to be moved, before moving them (see Fig. 8.2).
Then, when the cells are shifted to the new PCU, the BVC RESET and UNBLOCK
procedures (see 7.3.1.1) are performed for the same cells (see Fig. 8.3).
In the Siemens-BSS implementation, neither the Cell Identifier nor the BVCI of the
moved cells change.

178 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

PCU:0 BSSGP:BVC_BLOCK_34
PTPPKF
NSEI:0
RA=2 BSSGP:BVC_BLOCK_ACK_34
old_PCU_load=9/3=3 PTPPKF
1 PDCH RA=1
new_PCU_load=6/3=2
BVC=4 3 PDCH
BVC=23 PCU_TS_Gb= 3
PTPPKF
RA=2 PTPPKF PTPPKF
1 PDCH RA=2 RA=1
BVC=24 1 PDCH 3 PDCH
BVC=33 BVC=34

PCU:2
NSEI:2
PCU_load=4/2=2

PCU_TS_Gb= 2 SGSN
New PCU is created

PCU:4
NSEI:4 BSSGP:BVC_BLOCK_77
old_PCU_load=3/1=3 PTPPKF
new_PCU_load=2/1=2 RA=3 BSSGP:BVC_BLOCK_ACK_77
1 PDCH
BVC=77 PCU_TS_Gb= 1
PTPPKF
RA=5
2 PDCH
BVC=35

Fig. 8.2 Example of PTPPKF Distribution when a New PCU is Created - Step 1

A30808-X3247-L24-5-7618 179
GPRS/EGPRS Global Description Information
System

PCU:0
PTPPKF
NSEI:0
RA=2 PTPPKF
PCU_load=6/3=2
1 PDCH RA=1
BVC=4 3 PDCH
BVC=23 PCU_TS_Gb= 3
PTPPKF
RA=2 PTPPKF
1 PDCH RA=2
BVC=24 1 PDCH
BVC=33

PCU:2 BSSGP:BVC_RESET_34:Cell Identifier


NSEI:2
BSSGP:BVC_RESET_ACK_34
PCU_load=4/2=2 PTPPKF
RA=1 BSSGP:BVC_UNBLOCK_34
3 PDCH
BVC=34 BSSGP:BVC_UNBLOCK_ACK_34 SGSN

PCU_TS_Gb= 2

PTPPKF BSSGP:BVC_RESET_77:Cell Identifier


RA=3 BSSGP:BVC_RESET_ACK_77
1 PDCH
BVC=77 BSSGP:BVC_UNBLOCK_77
BSSGP:BVC_UNBLOCK_ACK_77

PCU:4
NSEI:4
PCU_load=2/1=2

PCU_TS_Gb= 1
PTPPKF
RA=5
2 PDCH
BVC=35

Fig. 8.3 Example of PTPPKF Distribution when a New PCU is Created - Step 2

180 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

8.1.3 PCU Crash


When a PCU crashes, the algorithm redistributes the configured PTPPKFs among the
remaining PCUs.
Since the PCU is crashed, it is not possible to perform the BVC BLOCK procedure on
the damaged PCU (in any case the SGSN can detect this anomaly because the L2 layer
is nolonger working).
The PTPPKFs are moved onto healthy PCUs, and then a BVC RESET procedure is
started for each shifted PTPPKF object (see Fig. 8.4).
The routing area information is not considered by the algorithm, to save the time of the
elaboration.

PCU:0
PTPPKF
NSEI:0 BSSGP:BVC_RESET_34:Cell Identifier
RA=2 PTPPKF
old_PCU_load=6/3=2
1 PDCH RA=1
new_PCU_load=9/3=3 BSSGP:BVC_RESET_ACK_34
BVC=4 3 PDCH
BVC=23 PCU_TS_Gb= 3
PTPPKF
RA=2 PTPPKF PTPPKF BSSGP:BVC_UNBLOCK_34
1 PDCH RA=2 RA=1 BSSGP:BVC_UNBLOCK_ACK_34
BVC=24 1 PDCH 3 PDCH
BVC=33 BVC=34

PCU:2 NO block procedures


NSEI:2 on Gb interface.
FAILED PTPPKF The PCU fails suddenly
RA=1
3 PDCH
PCU_TS_Gb= 2
BVC=34 SGSN
PTPPKF
RA=3
1 PDCH
BVC=77

PCU:4 BSSGP:BVC_RESET_77:Cell Identifier


NSEI:4 PTPPKF
old_PCU_load=2/1=2 RA=3 BSSGP:BVC_RESET_ACK_77
new_PCU_load=3/1=3 1 PDCH
BVC=77
PCU_TS_Gb= 1
PTPPKF
RA=5
2 PDCH BSSGP:BVC_UNBLOCK_77
BVC=35 BSSGP:BVC_UNBLOCK_ACK_77

Fig. 8.4 Example of PTPPKF Distribution in Case of PCU Crash

A30808-X3247-L24-5-7618 181
GPRS/EGPRS Global Description Information
System

8.1.4 PCU Comes Back in Service


When the PCU comes back in service, the PTPPKFs related to this PCU, before the
fault, are put back on it (see Fig. 8.5).
This process is not considered as a PTPPKF reconfiguration, in fact in order to mini-
mize the TBFs release, the reallocation algorithm is not performed when a PCU comes
back in service after a fault, but the PTPPKFs belonging to this PCU will be put back on
it.
Eventually, the PTPPKFs that belonged to the other faulty PCUs are involved in this
reallocation procedure.
There are two exceptions for this case:
1. if a PCU comes back in service, but before another PCU has come in service for the
first time (i.e. it was created), then the algorithm performs a total reconfiguration;
2. if some modifications regarding the FRLs allocation of the PCU has been done
during the period of time the PCU/Gb was down, then the algorithm performs a total
reconfiguration.
In fact there is a difference between putting back PTPPKFs to the PCU they belonged
before and a total redistribution itself.
In the first case, the previously moved cells are returned back without a new load calcu-
lation; the reason for not doing so is that in this case a new recalculation is simply due
to the need of managing short downs of Gb or PCU: it makes no sense to redistribute
everything again and again simply because, for example, a PCM line (containing the last
FRL of a PCU) is not stable.

182 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

PCU:0 PTPPKF BSSGP:BVC_BLOCK_34


NSEI:0 RA=2
old_PCU_load=9/3=3 PTPPKF BSSGP:BVC_BLOCK_ACK_34
1 PDCH RA=1
new_PCU_load=6/3=2 BVC=4 3 PDCH
BVC=23 PCU_TS_Gb= 3
PTPPKF
RA=2 PTPPKF PTPPKF
1 PDCH RA=2 RA=1
BVC=24 1 PDCH 3 PDCH
BVC=33 BVC=34

PCU:2 BSSGP:BVC_RESET_34:Cell Identifier


NSEI:2
BSSGP:BVC_RESET_ACK_34
IN-SERVICE PTPPKF
RA=1 BSSGP:BVC_UNBLOCK_34
3 PDCH
BVC=34 BSSGP:BVC_UNBLOCK_ACK_34 SGSN

PCU_TS_Gb= 2

PTPPKF BSSGP:BVC_RESET_77:Cell Identifier


RA=3 BSSGP:BVC_RESET_ACK_77
1 PDCH
BVC=77 BSSGP:BVC_UNBLOCK_77
BSSGP:BVC_UNBLOCK_ACK_77

PCU:4
NSEI:4 PTPPKF BSSGP:BVC_BLOCK_77
old_PCU_load=3/1=3 RA=3
1 PDCH BSSGP:BVC_BLOCK_ACK_77
new_PCU_load=2/1=2
BVC=77

PTPPKF
RA=5
2 PDCH
BVC=35

Fig. 8.5 Example of PTPPKF Distribution when a PCU Comes Back in Service

A30808-X3247-L24-5-7618 183
GPRS/EGPRS Global Description Information
System

8.1.5 Time Needed to Execute PTPPKF Reconfiguration


For what concerns the PTPPKFs distribution/redistribution algorithm, to avoid oscillation
in cases of fast PCU changes, a mechanism based on a Guard Timer is provided.
This timer lasts 5 seconds and it starts each time there is a modification in Gb/PCU
status. When the timer runs no calculations are executed to redistribute GPRS/EGPRS
cells (even if there are some changes in Gb/PCU status).
Obviously, this timer does not regard newly created PTPPKFs, because in this case
there is not a balancing procedure, but the created PTPPKF is simply put on the less
loaded PCU available in that moment.
In any case, the calculation algorithm is always executed in a few milliseconds, apart
fromsome calculations taking into account routing area considerations.
Calculations taking into account routing area considerations lasts some tens of millisec-
onds, so they cannot be done during normal TDPC working. Thus they are done only at
init phase (see "8.1.1 System Initialization").
The redistribution procedure takes, in general, some seconds. The worst case concerns
the moving of a PTPPKF from one PCU to another one: the PTPPKF needs to be
blocked and internally deleted on the old PCU, then created and unblocked on the new
one. This process takes about 7 seconds in total. This time, added to the 5 seconds of
the guard timer, makes a total of 12 seconds that are needed to redistributes PTPPKFs
to a PCU going into service.
This time is more or less independent from the number of involved PTPPKFs being done
in burst parallel on the various PTPPKFs. There are no significant differences between
the handling of PPCU and PPXU boards.

8.2 PCU Overload Management


For what concerns the PTPPKFs distribution/redistribution algorithm, no reconfiguration
is foreseen due to Overload reasons, since as it has been described in 8.1, the only trig-
gers are Down/Restart of PCU/Gb.
The overload prevention mechanism is based on the real time of the card.
The PCU overload management alarm is started when PCU processor real time
exceeds a threshold. The operating system measures the PCU processor real time. A
PCU cyclic task checks whether the percentage of the real time is greater than an upper,
non-configurable, threshold; if this is true, the cyclic task sends a message indicating
that the machine is overloaded to the GPRS/EGPRS overload handler process.
The cyclic task stops the overload message sending to the GPRS/EGPRS overload
handler when the percentage of the real time is smaller than a lower, non-configurable,
threshold.
The GPRS/EGPRS overload handler acts structurally as the already existing BSC and
BTS overload handlers.
At the very first signal of PCU Overload, an alarm is emitted towards the LMT/RC. The
overload process is controlled by BSCT17 and BSCT18 timers and it is based upon the
progressive application of protection actions, if the overload situation persists.
If the overload situation disappears these protection actions are progressively removed.
The protection action, used by the GPRS/EGPRS overload handler to reduce the PS
service traffic, is based upon the sending of System Information indicating that
GPRS/EGPRS services are not present in the cell. This is done for a progressively

184 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

increasing number of cells (at steps of 10 cells) allocated on the involved PCU, if the
overload situation persists.
On the contrary, if the overload finishes, System Information indicating that
GPRS/EGPRS service is present again are sent; this is done in steps of 5 cells that were
formerly "GPRS/EGPRS barred". When all cells are in the original situation, the PCU
overload alarm is ceased.

A30808-X3247-L24-5-7618 185
GPRS/EGPRS Global Description Information
System

9 GPRS/EGPRS Procedures

9.1 Mobile Stations for Packet Switched Services


A GPRS/EGPRS MS can operate in one of three modes of operation. The mode of oper-
ation depends on:
– the services that the MS is attached to, i.e., only PS services or both PS and CS
services;
– the MS capabilities to operate GPRS/EGPRS and other GSM services simulta-
neously.
The three modes of operation are:
• Class-A mode of operation: the MS is attached to both GPRS/EGPRS and other
GSM services, and the MS supports simultaneous operation of GPRS/EGPRS and
other GSM services;
• Class-B mode of operation: the MS is attached to both GPRS/EGPRS and other
GSM services, but the MS can only operate one set of services at a time. In network
operation mode III (see "9.8.3.1 Network Operation Modes for Paging"), a MS that
is capable of monitoring only one paging channel at a time cannot operate in class
B mode of operation. In this case, such a MS reverts to class-C mode of operation.
• Class-C mode of operation: the MS is exclusively attached to GPRS/EGPRS
services.
For what concerns EGPRS mobiles only, a supplementary distinction exists. Two mobile
classes are foreseen:
– the first class of mobiles, which allows a cost efficient and fast implementation,
supports the 8PSK modulation in downlink direction and the GMSK modulation in
uplink direction;
– the second class of mobiles has a more advanced capability, supporting the 8PSK
modulation in both uplink and downlink directions.

9.2 Network Structure


On the air interface, the network organization structure remains the same as in the GSM
implementation. An additional identifier is introduced to group cells that support packet
switched services in the Location Area (LA).
This identifier is named Routing Area (RA) and it is a sub entity of the Location Area.
The Routing Area is a more precise description of the current position of the
GPRS/EGPRS mobile station; one LA can include up to 256 RAs.
Routing Area numbering is not unique in the network, but it is unique in the Location
Area domain; to identify a Routing Area two parameters have been introduced:
• the Routing Area Code (RACODE): it identifies univocally the routing area inside the
location area;
• the Routing Area Color (RACOL): it allows the Mobile Station to distinguish between
two routing areas that have the same Routing Area Code, but belong to two different
Location Areas.
Fig. 9.1 shows an example of network structure.

186 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Location Area

Routing Area

Fig. 9.1 Network Structure: Example

9.3 Mobility Management Functionalities

9.3.1 Mobility Management States


The Mobility Management (MM) activities related to a GPRS/EGPRS subscriber are
characterized by one of three different MM states (see Fig. 9.2). Each state describes
a certain level of functionality and allocated information. The information, which is held
on both MS and SGSN sides, is denoted MM context.

A30808-X3247-L24-5-7618 187
GPRS/EGPRS Global Description Information
System

Fig. 9.2 Mobility Management States

9.3.1.1 IDLE State


In GPRS IDLE state the subscriber is not attached to the GPRS/EGPRS mobility
management. Both the MS and the SGSN contexts hold no valid location or routing
information for the subscriber. The subscriber-related mobility management procedures
are not performed.
PLMN selection and GPRS/EGPRS cell selection and re-selection processes (see 10.1)
are performed by the MS.
In order to establish MM contexts in the MS and in the SGSN, the MS will perform the
GPRS Attach procedure (see "9.3.2.1 Attach Function").
When the GPRS attach procedure has been executed, the MM state moves from IDLE
to READY: the MS requests access and a logical link to an SGSN is established.

9.3.1.2 STAND-BY State


In STANDBY state the subscriber is attached to GPRS/EGPRS mobility management.
The MS and the SGSN have established MM contexts for the subscriber.
Pages for PTP data transfers or signaling information transfers may be received. It is
also possible to receive pages for CS services via the SGSN (only if the Gs interface
between the MSC and the SGSN is implemented). PTP data reception and transmission
are not possible in stand-by state.
The MS performs Routing Area updates and GPRS/EGPRS cell selection and re-selec-
tion locally. The MS executes mobility management procedures to inform the SGSN
when it has entered a new RA. The MS does not inform the SGSN on a change of cell

188 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

in the same RA. Therefore, for MSs in STANDBY state, the location information in the
SGSN MM context contains only the routing area identity (RAI).
The Routing Area Identity is defined as:
i RAI = MCC + MNC + LAC + RAC
where:
- MCC = mobile country code
- MNC = mobile network code
- LAC = location area code
- RAC = routing area code

The MS may initiate activation or deactivation of PDP contexts while it is in STANDBY


state. A PDP context must be activated before data can be transmitted or received for
this PDP context.
The SGSN may have to send data or signalling information to a MS in STANDBY state;
in this case the SGSN sends a Paging Request in the routing area where the MS is
located. The MM state in the MS changes to READY when the MS responds to the page,
and in the SGSN when the response to paging is received. The MM state in the MS also
changes to READY when data or signaling information is sent from the MS and the MM
state in the SGSN changes to READY when data or signalling information is received
from the MS.
The MS or the network may initiate the GPRS Detach procedure to move to the IDLE
state. After expiry of the mobile reachable timer, the SGSN may perform an implicit
detach in order to return the MM contexts in the SGSN to IDLE state. The MM and PDP
contexts may then be deleted.
In particular, the following procedures cause the transition from the standby state to the
other MM states:
• moving from STANDBY to IDLE:
– Implicit Detach: the MM and PDP contexts in the SGSN return to IDLE and INAC-
TIVE state;
– Cancel Location: the SGSN receives a Cancel Location message from the HLR,
and removes the MM and PDP contexts;
• moving from STANDBY to READY:
– PDU transmission: the MS sends a LLC PDU to the SGSN, possibly in response
to a page;
– PDU reception: the SGSN receives a LLC PDU from the MS.

9.3.1.3 READY State


In READY state the MM context corresponds to the STANDBY MM context extended by
location information for the subscriber on the cell level. The MS performs mobility
management procedures to provide the network with the actual selected cell.
GPRS/EGPRS cell selection and re-selection is done locally by the MS, or optionally
may be controlled by the network (see "10.3 Network Controlled Cell Reselection and
Traffic Control Management").
The MS may send and receive PDP PDUs in this state. The network does not initiate PS
pages for a MS in READY state; pages for other services may be done via the SGSN.
The MS may activate or deactivate PDP contexts while in READY state. The READY
state is supervised by a timer. A MM context moves from READY state to STANDBY
state when the READY timer expires.

A30808-X3247-L24-5-7618 189
GPRS/EGPRS Global Description Information
System

In order to move from READY state to IDLE state, the MS initiates the GPRS Detach
procedure.
The following procedures cause the transition from the ready state to the other MM
states:
• moving from READY to STANDBY:
– READY timer expiry (see " READY TIMER"): the MS and the SGSN MM contexts
return to STANDBY state;
– Force to STANDBY: the SGSN indicates an immediate return to STANDBY state
before the READY timer expires;
– abnormal RLC condition: the SGSN MM context returns to STANDBY state in
case of delivery problems on the radio interface;
• moving from READY to IDLE:
– GPRS Detach: the MS or the network request that the MM contexts return to IDLE
state and that the PDP contexts return to INACTIVE state. The SGSN may delete
the MM and PDP contexts. The PDP contexts in the GGSN are deleted.
– Cancel Location: the SGSN receives a Cancel Location message from the HLR
and removes the MM and PDP contexts.

READY TIMER
The READY timer controls the time that a MS remains in READY state before going to
the STANDY state. In the MS the READY timer is reset and begins running when a LLC
PDU is transmitted; in the SGSN the timer begins running when a LLC PDU is correctly
received. When the READY timer expires, both the MS and SGSN MM contexts return
to STANDBY state. The length of the READY timer is the same in the MS and in the
SGSN. If the READY timer length is set to zero, the MS is immediately forced into
STANDBY state.

9.3.2 Mobility Management Procedures


MM procedures use the LLC and RLC/MAC protocols for message transmission across
the Um interface. The MM procedures provide information to the underlying layers to
enable reliable transmission of MM messages on the Um interface. User data can be
transmitted during MM signalling procedures.
User data transmitted during attach, authentication, and routing area update procedures
may be lost and therefore may have to be retransmitted. In order to minimize the need
for retransmission, the MS and the SGSN do not transmit user data during the attach,
authentication, and routing area update procedures.

9.3.2.1 Attach Function


The GPRS/EGPRS attach procedure is executed by the MS. In the attach procedure,
the MS provides its identity and an indication of which type of attach it wants to execute.
Two different types of attach are foreseen (both of them executed towards the SGSN):
– the GPRS attach;
– the combined GPRS/IMSI attach.
The combined attach allows the MS to register itself both in the SGSN and in the MSC,
but this procedure can be executed only when the network works in a co-ordinated way,

190 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

i.e., only when the Gs interface between the MSC and the SGSN is configured (see also
"9.8.3.1 Network Operation Modes for Paging").
If the network operates in Mode I (see "9.8.3.1 Network Operation Modes for Paging"),
i then a MS that is both GPRS/EGPRS-attached and IMSI-attached performs the
combined RA/LA update procedure.
If the network operates in Mode II or III, then a GPRS/EGPRS-attached MS, that has
the capability to be simultaneously GPRS/EGPRS attached and IMSI-attached,
performs the (not-combined) Routing Area Update procedure, and accesses the CCCH
channel for CS operation.

After having executed the GPRS attach procedure, the MS is in READY state and MM
contexts are established in both the MS and the SGSN. The MS may then activate PDP
contexts as described in "9.7 Activation and Deactivation of a PDP Context".
An IMSI-attached MS that can only operate in class-C mode of operation follows the
normal IMSI detach procedure before making a GPRS attach. A GPRS-attached MS in
class-C mode of operation will always perform a GPRS detach before making an IMSI
attach.
The steps of the Attach procedure are illustrated below:
1. The MS initiates the attach procedure by the transmission of the Attach Request
message to the SGSN. The message contains the following information:
– IMSI or P-TMSI: IMSI is included if the MS does not have a valid P-TMSI. If the
MS has a valid P-TMSI, then P-TMSI and the old RAI associated with P-TMSI are
included;
– Classmark: it contains the MS's multislot capabilities and supported ciphering
algorithms for PS services;
– Attach Type: it indicates which type of attach that must be performed, i.e.,
GPRS/EGPRS attach only, GPRS/EGPRS Attach while already IMSI attached, or
combined (E)GPRS/IMSI attach;
– DRX Parameters: indicate when the MS is in a non-sleep mode and when it is
able to receive paging requests and channel assignments (see
"9.8.3.2 Discontinuous Reception").
2. The SGSN sends the Attach Accept message to the MS; P-TMSI is included if the
SGSN allocates a new P-TMSI;
3. If P-TMSI has been changed, the MS acknowledges the received P-TMSI by the
Attach Complete message.
P-TMSI (Packet Temporary Mobile Subscriber Identity) is optionally sent by the SGSN
i to the MS in Attach Accept and Routing Area Update Accept messages. If the P-TMSI
signature has been sent by the SGSN to the MS because a new P-TMSI has been allo-
cated by the SGSN, then the MS includes the received P-TMSI signature in the next
Routing Area Update Request or in the next Attach Request for identification checking
purposes. In both the Attach and Routing Area Update procedures, the SGSN
compares the P-TMSI signature sent by the MS with the signature stored in the SGSN.
If the values do not match, the SGSN should use the security functions to authenticate
the MS. The P-TMSI signature parameter has only local significance in the SGSN that
allocated the signature.
P-TMSI is similar the GSM T-IMSI (temporary IMSI).

A30808-X3247-L24-5-7618 191
GPRS/EGPRS Global Description Information
System

If the Attach Request cannot be accepted, the SGSN returns the Attach Reject
message to the MS. The message contains the cause that has generated the rejection.
A GPRS/EGPRS-attached MS makes IMSI attach via the SGSN with the combined
RA/LA update procedure if the network operation mode is I. In network operation modes
II and III, or if the MS is not GPRS/EGPRS-attached, the MS makes IMSI attach as
defined in GSM.

9.3.2.2 Detach Function


The Detach function allows a MS to inform the network that it wants to make a
GPRS/EGPRS and/or IMSI detach, and allows the network to inform a MS that it has
been GPRS/EGPRS-detached or IMSI-detached by the network.
A GPRS/EGPRS-attached MS sends a Detach Request message to the SGSN, indi-
cating a GPRS/EGPRS detach.
The Detach Request message contains an indication to determine if the detach is due
to switch off or not. The indication is needed to determine whether or not a Detach
Accept message will be returned.

9.4 Radio Resource Management


Radio Resource (RR) management procedures are characterized by two different RR
operating modes (see Fig. 9.3). Each mode describes a certain amount of functional-
ities and information allocated.

- No RLC context in
Packet Idle State
MS & SGSN

- Mobile originated
call - Deactivation of RLC context
- Answer to paging

Packet Transfer - RLC context in MS & SGSN


- Routing context between
State MS & SGSN

Fig. 9.3 Radio Resource States

9.4.1 Packet Idle State


In packet idle mode no Temporary Block Flows exist (see "4.3 Temporary Block Flow").
Upper layers can require the transfer of LLC PDUs which, implicitly, may trigger the
establishment of a TBF and the transition to packet transfer mode (see Fig. 9.3).
In packet idle mode the MS listens to the PBCCH and to the paging sub-channel for the
paging group to which the MS belongs. If the PCCCH is not present in the cell, the
mobile station listens to the BCCH and to the relevant paging sub-channels.

192 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

9.4.2 Packet Transfer State


In packet transfer mode, the mobile station uses the allocated radio resources to
transmit radio blocks. Continuous transfer of one or more LLC PDUs is possible.
Concurrent TBFs may be established in opposite directions. Transfer of LLC PDUs in
either RLC acknowledged or RLC unacknowledged mode is provided.
When selecting a new cell, a mobile station:
1. leaves the packet transfer mode
2. enters the packet idle mode
3. switches to the new cell
4. reads the system information of the new cell
5. resumes the packet transfer mode in the new cell.
While operating in packet transfer mode, a mobile station belonging to GPRS/EGPRS
class A may simultaneously enter the CS dedicated (connected) mode. A mobile station
belonging to GPRS/EGPRS class B leaves the packet transfer modes before entering
the CS dedicated mode.

9.5 Correspondence between RR States and MM States


Tab. 9.1 provides the correspondence between Radio Resource states and Mobility
Management states.

MM States READY STAND-BY


RR States Packet Transfer Packet Idle state Packet Idle state
state

Tab. 9.1 Correspondence between MM States and RR States

9.6 Packet Data Protocol Functionalities


The subscription of a point to point PS service contains one or more PDP addresses.
Each PDP address is described by an individual PDP context in the MS, in the SGSN,
and in the GGSN.
Every PDP context exists independently in one of two PDP states. The PDP state indi-
cates whether or not the PDP address is activated for data transfer. Activation and deac-
tivation procedures are described in section 9.7. All PDP contexts of a subscriber are
associated with the same MM context for the IMSI of that subscriber.

A30808-X3247-L24-5-7618 193
GPRS/EGPRS Global Description Information
System

Fig. 9.4 Packet Data Protocol States

9.6.1 INACTIVE State


The INACTIVE state characterizes the data service for a certain PDP address of the
subscriber as not activated. The PDP context contains no routing or mapping informa-
tion to process PDUs related to that PDP address. No data can be transferred.
A changing location of a subscriber causes no update for the PDP context in INACTIVE
state even if the subscriber is attached to the PS MM.
If the GGSN is allowed to initiate the activation of the PDP context for that PDP address,
mobile-terminated PTP packets, received in INACTIVE state by the GGSN, may initiate
the Network-Requested PDP Context Activation procedure. Otherwise, mobile-termi-
nated PTP packets received in INACTIVE state invoke error procedures in the GGSN
(for instance an IP packet is discarded). Other error procedures may be introduced on
the application level, but this is outside of the scope of the current document.
The MS initiates the transition from the INACTIVE state to the ACTIVE state by initiating
the PDP Context Activation procedure (see "9.7.1 PDP Context Activation").

9.6.2 ACTIVE State


In ACTIVE state, the PDP context for a specific PDP address is activated in the MS, in
the SGSN and in the GGSN. The PDP context contains mapping and routing information
for transferring PDUs (for that particular PDP address) between the MS and the GGSN.
The ACTIVE PDP state is permitted only when the mobility management state of the
subscriber is STANDBY or READY.

194 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

An active PDP context for a MS is moved to the INACTIVE state when the deactivation
procedure is initiated. All active PDP contexts for a MS are moved to the INACTIVE state
when the MM state changes to IDLE.

9.7 Activation and Deactivation of a PDP Context


These functions are only meaningful at the NSS level and in the MS, and do not directly
involve the BSS. A MS in STANDBY or READY state can initiate these functions at any
time to activate or deactivate a PDP context in the MS, in the SGSN, and in the GGSN.
Upon receiving an Activate PDP Context Request message, the SGSN initiates proce-
dures to set up PDP contexts.
Upon receiving a Deactivate PDP Context Request message, the SGSN initiates proce-
dures to deactivate PDP contexts.

9.7.1 PDP Context Activation


The PDP context activation procedure is used by the mobile station to obtain an IP
address from the network, and to negotiate service parameters such as delay class
(average and peak) and throughput (average and peak). Currently the network foresees
only the ‘best effort’ quality of service, giving the available resources to the MS, without
taking into account any mobile request.
In the following, the PDP Context Activation procedure is described:
1. In order to activate a PDP context, the mobile station sends a CHANNEL REQUEST
message to the network. The network, after having reserved a channel on the BTS,
sends an IMMEDIATE ASSIGNMENT message carrying the PACKET UPLINK
ASSIGNMENT information element. This message reserves an uplink resource (a
time slot, with TFI and USF) to the mobile station, allowing it to transmit the ACTI-
VATE PDP CONTEXT REQUEST;
2. The MS sends the ACTIVATE PDP CONTEXT REQUEST message to the SGSN.
The following information is contained in the message:
– PDP Address: used by the MS to indicate whether it requires the use of a static
PDP address or whether it requires the use of a dynamic PDP address; the MS
will leave the PDP Address empty to request a dynamic PDP address;
– Access Point Name: it is a logical name referring to the external packet data
network to which the subscriber wishes to connect;
– QoS: indicates the desired QoS profile;
3. Security functions are executed by the SGSN;
4. The SGSN validates the ACTIVATE PDP CONTEXT REQUEST using PDP Type,
PDP Address, and Access Point Name provided by the MS and the PDP context
subscription records;
5. The SGSN sends a CREATE PDP CONTEXT REQUEST (PDP Type, PDP Address,
Access Point Name, Negotiated QoS) message to the affected GGSN;
6. The GGSN uses the Access Point Name to find an external network. The GGSN
creates a new entry in its PDP context table and generates a Charging ID. The new
entry allows the GGSN to route PDP PDUs between the SGSN and the external
PDP network, and to start charging;
7. The GGSN then returns a CREATE PDP CONTEXT RESPONSE message to the
SGSN. PDP Address is included if the GGSN has allocated a PDP address;

A30808-X3247-L24-5-7618 195
GPRS/EGPRS Global Description Information
System

8. The SGSN selects the Radio Priority based on the Negotiated QoS, and returns the
ACTIVATE PDP CONTEXT ACCEPT message to the MS.
The RLC/MAC layer supports four Radio Priority levels and an additional level for signal-
i ling messages as defined in GSM 04.60. Upon uplink access, the MS can indicate one
of the four priority levels, and whether the cause for the uplink access is user data or
signalling message transmission. This information is used by the BSS to determine the
radio access precedence (i.e., access priority) and the service precedence (i.e., transfer
priority under congested situation).
The Radio Priority concept is related to the QoS one, i.e., a higher Quality of Service
corresponds to a higher Radio Priority. The Radio Priority is coded as follows:
- 0: Radio Priority 1 (Highest priority, used also for signalling)
- 1: Radio Priority 2
- 2: Radio Priority 3
- 3: Radio Priority 4 (Lower priority)

9.7.2 PDP Context Deactivation


The PDP Context Deactivation procedure is executed with the following steps:
1. The MS sends a DEACTIVATE PDP CONTEXT REQUEST message to the SGSN;
2. Security functions are executed by the SGSN;
3. The SGSN sends a DELETE PDP CONTEXT REQUEST message to the GGSN.
The GGSN removes the PDP context and returns a DELETE PDP CONTEXT
RESPONSE message to the SGSN. If the MS used a dynamic PDP address, then
the GGSN releases this PDP address and makes it available for subsequent activa-
tions by other MSs;
4. The SGSN returns a DEACTIVATE PDP CONTEXT ACCEPT message to the MS.
At GPRS/EGPRS detach, all PDP contexts for the MS are implicitly deactivated.
i

9.8 Access to the Network (Establishment of a TBF)


The establishment of a Temporary Block Flow (TBF) is initiated to transfer LLC PDUs
between the network and the MS. The request to establish a TBF can be done:
• on CCCH if the PBCCH/PCCCH is not configured in the cell
• on PCCCH if the PBCCH/PCCCH is allocated in the cell
Two types of establishment exists:
– TBF establishment initiated by the MS, i.e., a MS initiated packet transfer (see 9.8.2)
– TBF establishment initiated by the network, i.e., a network initiated packet transfer
(see 9.8.3).

9.8.1 Medium Access Modes


Two types of access modes exist:
a) Dynamic Allocation: the assignment message includes the list of PDCHs and the
corresponding USF value for each assigned PDCH. A unique TFI is allocated and is
thereafter included in each RLC Data and Control Block related to that TBF. Dynamic
allocation is characterized by the MS that monitors the USF values on the allocated
PDCHs and transmits Radio blocks on those that currently bear the USF value
reserved for the usage of the MS;

196 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

b) Fixed Allocation: it is characterized by fixed allocation of radio blocks and PDCHs in


the assignment message, without assigning USF values. In Release BR7.0 Fixed
Allocation is not supported.

9.8.2 TBF Establishment Initiated by the MS on CCCH/PCCCH


The purpose of the packet access procedure is to establish a TBF to support the transfer
of LLC PDUs in the direction from the mobile station to the network. Packet access is
done on the PCCCH if it exists, otherwise packet access is done on the CCCH.
In the following, messages that are exchanged either on the CCCH or PCCCH channel
i are shown using the following method: messages exchanged on CCCH are normally
used, whereas the corresponding PCCCH messages are inserted in brackets.

The packet access procedure is initiated by the mobile station when a request to transfer
LLC PDUs comes from upper layers.
Two different access types exist:
• one phase access: the network responds reserving resources on the PDCH(s) to
allow uplink packet transfer of a number of Radio Blocks;
• two phase access: the network responds reserving resources for transmitting a
PACKET RESOURCE REQUEST message; this message is used by the MS to
better define the needed radio resources.
In both cases, when the uplink TBF is set up, the following parameters and radio
resources are allocated to the MS (with the assignment message):
– USF
– TFI
– Time Slot numbers
– Channel Coding Scheme
– ARFCN
– optionally, the Timing Advance parameters (TAI and Timeslot number); if the timing
advance index (TAI) is included in the uplink assignment construction, the mobile
station will use the continuous update timing advance mechanism, using PTCCH in
the same timeslot as the assigned PDCH (see 4.6). If a timing advance index (TAI)
field is not included, the continuous update timing advance mechanism will not be
used;
– MAC access mode (always set to dynamic in BR7.0, see "9.8.1 Medium Access
Modes").
In addition, the EGPRS uplink assignment contains the following information:
– the EGPRS modulation and coding scheme;
– the EGPRS window size;
– information on whether or not retransmitted uplink data blocks must be reseg-
mented.
When a MS tries to access to the network, the GPATH parameter indicates if the MS,
according to its priority class, is authorized to perform a random access to request
packet switched services.

9.8.2.1 8 Bit or 11 Bit Uplink Access


To access the GSM network, a slotted-aloha protocol is used; the access is performed
sending a traditional 8 bit Access Burst type.

A30808-X3247-L24-5-7618 197
GPRS/EGPRS Global Description Information
System

According to ETSI specifications, a new enhanced Access Burst type with 11 informa-
tion bit can be sent by the mobile station to try to access the network. In fact, 8 bit of
information does not allow to widely specify the needed resources. To overcome this
bottleneck, an access burst using 11 information bit is defined. Fig. 9.5 shows the
coding process of the 11 bit access burst.
Therefore, a GPRS/EGPRS mobile station can access the network by using an 8 bit or
an 11 bit access burst, in particular:
• the CHANNEL REQUEST message sent on RACH is always formatted with 8 bit of
information;
• the PACKET CHANNEL REQUEST sent on PRACH can be formatted either with 8
or 11 bit.
The possibility of using one message type or the other one for the PACKET
CHANNEL REQUEST depends on network settings: the capability of the network to
receive 8 or 11 bit length message is broadcast by the System Information param-
eter ABUTYP, that indicates the allowed type of access.
The ABUTYP parameter setting also indicates which type of access burst (8 or 11 bit)
must be sent:
– for PACKET CONTROL ACKNOWLEDGMENT messages, that are formatted as
four access burst;
– in the PTCCH channel, for the continuous timing advance estimation (see 4.6).
Besides, the 11 bit access burst is the only one supporting the EGPRS PACKET
i CHANNEL REQUEST message, that can be used from EDGE mobile stations to access
a cell, see "9.8.2.4 TBF Establishment for EDGE Mobile Stations".

The advantages of the 11 bit uplink access are the following:


– it more often allows a one phase access instead of a two phases access;
– it speeds up the call set-up and decreases the signalling load.
The 8 bit access on PRACH or RACH is used in case of PAGING RESPONSE, CELL
UPDATE, MM PROCEDUREs and in all cases that the MS requires sending no more
information than the MS class and priority.
The 11 bit access on PRACH is used in all cases described for 8 bit access, but with
additional information to be carried in the access phase, e.g., an enhanced random
reference number leading to less probability of MS collision when trying to establish an
uplink TBF.

11 information bit
1/2
+ 6 bit 36 encrypted
Convolutional
6 parity bit puncturing bit
coder
+
4 tail bit

Access Burst
Tail Synchronization Sequence Information bit Tail Guard period

36 bit

Fig. 9.5 Coding of the 11 Bit Access Burst

198 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

9.8.2.2 Establishment using a One Phase Access


A mobile station initiates the packet access procedure by scheduling the sending of
CHANNEL REQUEST (PACKET CHANNEL REQUEST) messages on the RACH
(PRACH), and simultaneously leaving the packet idle mode.
The topics described in this chapter are valid for GPRS MSs and also for EDGE MSs
i when the EGPRS PACKET CHANNEL REQUEST is not supported in the cell (see
"9.8.2.4 TBF Establishment for EDGE Mobile Stations").

Then, the mobile station monitors the full CCCH (PCCCH) corresponding to its RACH
(PRACH) to wait for the network answer.
If the PCCCH is configured and the mobile station receives the PERSISTENCE_LEVEL
i parameter from the network, the value of the PERSISTENCE_LEVEL parameter is
taken into account at the next PACKET CHANNEL REQUEST attempt (see
"9.8.2.6 Uplink Access on PRACH (Access Persistence Control)").

When the MS has sent the CHANNEL REQUEST (PACKET CHANNEL REQUEST)
message, the following behaviors will occur, according to its class:
– a mobile station in class A or class B mode of operation will respond to a paging
message indicating a circuit switched connection establishment;
– a mobile station in class B mode of operation may abort the packet access proce-
dure, if it receives a paging message indicating the establishment of circuit switched
connections;
– mobile stations in class C mode of operation will not respond to any type of paging
messages during the packet access procedure.
CHANNEL REQUEST (PACKET CHANNEL REQUEST) messages are sent on the
RACH (PRACH) and contain, beside the indication of the type of access, the required
parameters to indicate the demand of radio resources from the MS.
When the network receives the CHANNEL REQUEST (PACKET CHANNEL
REQUEST) message, it may assign radio resources on one or more PDCHs, to be used
by the mobile station for the TBF.
The allocated PDTCH(s) and PACCH resources are assigned to the mobile station in
the IMMEDIATE ASSIGNMENT (PACKET UPLINK ASSIGNMENT) message, sent on
any AGCH (PAGCH) block on the same CCCH (PCCCH) on which the network has
received the CHANNEL REQUEST (PACKET CHANNEL REQUEST) message.
In the one phase access, the reservation is done according to the information about the
requested resources, comprised in the channel request.
On the RACH, in the CHANNEL REQUEST message, there are only 8 bit of information,
so there are only two available values for denoting PS calls; these values can be used
to request limited resources in the one phase access (EGPRS TBFs cannot be opened
using a one phase access on RACH, using the CHANNEL REQUEST message, see
"9.8.2.4 TBF Establishment for EDGE Mobile Stations") or to request a two phase
access.
On the PRACH, the PACKET CHANNEL REQUEST may contain more adequate infor-
mation about the requested resources and, consequently, uplink resources on one or
several PDCHs can be assigned by using the PACKET UPLINK ASSIGNMENT
message.
Fig. 9.6 shows a one phase access procedure on PCCCH.

A30808-X3247-L24-5-7618 199
GPRS/EGPRS Global Description Information
System

Fig. 9.6 One Phase Access on PCCCH

9.8.2.3 TBF Establishment using a Two Phases Access


In the first phase of a two phase access, the same procedure as for one phase access
is used, until the network sends the IMMEDIATE ASSIGNMENT (PACKET UPLINK
ASSIGNMENT) message. This message denotes a single block allocation, which indi-
cates to the MS the two phases access.
In this message, the network reserves a limited resource (single radio block) on one
PDCH to the mobile station, where the mobile station transmits a PACKET RESOURCE
REQUEST message.
The two phase access can be initiated either by the MS or the network, following these
rules:
• if PCCCH is provided in the cell, a two phase access can be initiated:
– by the network by ordering the mobile station to send a PACKET RESOURCE
REQUEST message. The order is sent implicitly to the mobile station in the
PACKET UPLINK ASSIGNMENT message by including the Single Block Alloca-
tion structure;
– by a mobile station, by requiring a two phase access in the PACKET CHANNEL
REQUEST message. In this case, if access is granted, the network will order the
mobile station to send a PACKET RESOURCE REQUEST message. The order
is sent implicitly to the mobile station in the PACKET UPLINK ASSIGNMENT
message by including the Single Block Allocation Structure.

200 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

• if no PCCCH is provided in the cell, a two phase access can be only initiated by a
mobile station, by requiring this type of access within the CHANNEL REQUEST
message.
When the network receives the PACKET RESOURCE REQUEST message, it responds
by sending either a PACKET UPLINK ASSIGNMENT message (radio resources assign-
ment on one or more PDCHs to be used by the mobile station for the TBF) or a PACKET
ACCESS REJECT message to the MS.
Fig. 9.7 shows a two phases access procedure on CCCH.

Fig. 9.7 Two Phases Access on CCCH


n the current release an improvement of the GAP between the Assignment and Packet
Resource Request (PRR) and/or the TBF start has been implemented. A gap of about
350, 450 ms has been detected between the IACMD and the PRR in case of two phase
access. For this reason a reduction and optimization of the overall delay for all the types
of PRR/TBF start has been applied for both idle channels and active channels. The
improvement has been achieved with the optimization of the commanded TBF starting
time.
In case of active channels it is not necessary the syncronization of the PDTs. As a
consequence the PRR/TBF start can be commanded without risk accordingly to the
internal RNLC timing.
In case of idle channels the Immediate Assignment (IACMD) is submitted by the PCU
parallel to the ongoing channel synchronization roughly before the ending of the channel
synchronization itself. Considering that the channel synchronization has a final cycle
without any TA and FN changes and an additional virtual cycle should be taken into
account as reserve (60-100ms) the PRR delay or the TBF start could be improved by
about 200ms.

A30808-X3247-L24-5-7618 201
GPRS/EGPRS Global Description Information
System

9.8.2.4 TBF Establishment for EDGE Mobile Stations


As described in "4.4.1 Packet Broadcast Control Channel (PBCCH)", the MS knows that
EGPRS is available in the cell, reading the GPRS Cell Option IE in SI13 or in PSI1 and
finding the EGPRS_PACKET_CHANNEL_REQUEST support indication field.
The EGPRS_PACKET_CHANNEL_REQUEST support indication field, is represented
by one bit, and can assume the values 0 or 1.
• In the first case, an EGPRS capable MS, will use the following to access a cell:
– the EGPRS PACKET CHANNEL REQUEST message for EGPRS uplink TBF
establishment on the PRACH when there is a PBCCH in the cell;
– the EGPRS PACKET CHANNEL REQUEST message for EGPRS uplink TBF
establishment on the RACH when there is no PBCCH in the cell.
• In the second case, an EGPRS capable MS, will use the following to access a cell:
– a two phase access with PACKET CHANNEL REQUEST message on the
PRACH for uplink TBF establishment when there is a PBCCH in the cell;
– a two phase access with CHANNEL REQUEST message on the RACH when
there is no PBCCH in the cell.
The EGPRS PACKET CHANNEL REQUEST message is formatted as 11 bit of informa-
tion, and it is supported only by the EDGE Access Burst. The EDGE Access Burst uses
two different training sequences, TS1 and TS2, to allow a mobile station to signal to the
network, during the access phase, its uplink capability. In fact, as described in
"9.1 Mobile Stations for Packet Switched Services" two mobile classes are foreseen for
what concerns EGPRS mobile stations; therefore:
– when the mobile station is an EGPRS one, with 8PSK modulation capability in uplink
direction, it uses the EDGE access burst with TS1 training sequence;
– when the mobile station is an EGPRS one, with GMSK modulation capability in
uplink direction, it uses the EDGE access burst with TS2 training sequence;
Instead, a GPRS mobile station will use a “traditional” access burst with the training
sequence already used for GSM (TS GSM).
It is important to note that only EDGE CUs are able to manage access bursts containing
TS1 and TS2 (and as a consequence the EGPRS PACKET CHANNEL REQUEST
message), so the EGPRS PACKET CHANNEL REQUEST message will be used to
access a cell only if the BCCH carrier supports EDGE, i.e., (see "5.1.2 Enabling EGPRS
Service in the Cell"):
a) at least one EDGE CU is configured in the BTS equipment handling the cell
b) the TRXMD parameter of the BCCH carrier is set to EDGE
c) the TrxCapability of the BCCH carrier is set to EDGE
Besides, the ABUTYP parameter that indicates the format of the information field of
access bursts must be set to ACBU11BIT.
Therefore, if the SI13 (or the PSI1) indicates that the cell is EGPRS capable, and
EGPRS PACKET CHANNEL REQUEST on RACH (PRACH) is supported in the cell, an
EGPRS mobile station sends the 11 bit EGPRS PACKET CHANNEL REQUEST
message.
If the SI 13 (or the PSI1) indicates that the cell is EGPRS capable and EGPRS PACKET
CHANNEL REQUEST on RACH /PRACH) is not supported in the cell, the EGPRS
mobile station will use the CHANNEL REQUEST message on RACH (or PACKET
CHANNEL REQUEST message on PRACH) message and initiates a two phases
access request. In the following, the access procedures are described in more detail.

202 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

EGPRS Uplink TBF Establishment using a One Phase Access


Regarding this kind of Uplink TBF establishment, different cases exist:
a) EGPRS PACKET CHANNEL REQUEST is supported:
– if the PBCCH channel is not supported, the MS sends an EGPRS PACKET
CHANNEL REQUEST message on the RACH channel.
The procedure is similar to those described in "9.8.2.2 Establishment using a One
Phase Access": the BSC answers with the IMMEDIATE ASSIGNMENT message
for Uplink EGPRS TBF, and then it sends a Packet Uplink Assignment for Uplink
EGPRS TBF if the TBF is multislot;
– if the PBCCH channel is supported, the MS sends an EGPRS PACKET
CHANNEL REQUEST message on the PRACH channel.
The procedure is similar to those described in "9.8.2.2 Establishment using a One
Phase Access": the BSC answers with a Packet Uplink Assignment for Uplink
EGPRS TBF.
b) EGPRS PACKET CHANNEL REQUEST is NOT supported: in this case, the one
phase access does not allow establishing an EGPRS TBF; a GPRS TBF will be allo-
cated following the procedure described in "9.8.2.2 Establishment using a One
Phase Access".

EGPRS Uplink TBF Establishment using a Two Phase Access


The following different cases exist regarding this kind of Uplink TBF establishment:
a) EGPRS PACKET CHANNEL REQUEST is supported:
– if the PBCCH channel is not supported, the MS sends an EGPRS PACKET
CHANNEL REQUEST message on RACH channel.
The BSC answers with IMMEDIATE ASSIGNMENT message, then the MS sends
a Packet Resource Request message following the procedure described in
"9.8.2.3 TBF Establishment using a Two Phases Access";
– if the PBCCH channel is supported, the MS sends an EGPRS PACKET
CHANNEL REQUEST message on PRACH channel.
The BSC answers with a Packet Uplink Assignment message, then the MS sends
a Packet Resource Request message following the procedure described in
"9.8.2.3 TBF Establishment using a Two Phases Access";
b) EGPRS PACKET CHANNEL REQUEST is NOT supported:
– if the PBCCH channel is not supported, the MS sends a Channel Request (with
two phase indication) on RACH channel. The BSC answers with IMMEDIATE
ASSIGNMENT (single block), then the MS sends a Packet Resource Request
message following the procedure described in "9.8.2.3 TBF Establishment using
a Two Phases Access";
– if the PBCCH is supported, the MS sends a Packet Channel Request (with two
phase indication) on PRACH channel. The BSC answers with PACKET UPLINK
ASSIGNMENT (single block), then the MS sends a Packet Resource Request
message following the procedure described in "9.8.2.3 TBF Establishment using
a Two Phases Access";

9.8.2.5 Contention Resolution


Contention resolution is an important part of RLC/MAC protocol operations, especially
because one channel allocation can be used to transfer a number of LLC frames.
As defined in the previous chapters, there are two basic access possibilities, the one
phase and the two phase access.

A30808-X3247-L24-5-7618 203
GPRS/EGPRS Global Description Information
System

The two phase access is inherently immune from the possibility that two MSs can
perceive the same channel allocation as their own. Namely, the second access phase,
i.e., the Packet Resource Request, uniquely identifies the MS by its TLLI. The same
TLLI is included in the Packet Uplink Assignment/Packet Downlink Assignment and no
mistakes are possible.
The Temporary Logical Link Identity (TLLI) identifies a GPRS/EGPRS user inside the
i cell. The relationship between TLLI and IMSI is known only in the MS and in the SGSN.
TLLI is derived from the P-TMSI allocated by the SGSN, or it is built by the MS.
The P-TMSI identifies the MS for location purposes, whereas TLLI identifies the MS
from the packet data transfer point of view.

The one phase access is somewhat insecure, and an efficient contention resolution
mechanism must be introduced.
The first part of the solution is the identification of the MS. The identification of transmit-
ting MS on the RLC/MAC level is necessary not only for contention resolution, but also
to be able to establish the RLC protocol entity for that Temporary Block Flow on the
network side. Additionally, TLLI is necessary to be able to match simultaneous uplink
and downlink packet transfers by taking into consideration multislot capability of that
MS. In order to uniquely identify the MS when sending on uplink, the RLC Header for all
of the RLC Data Blocks on the uplink is extended to include the TLLI, until the contention
resolution is completed on the MS side.
The second part of the solution is the notification from the network side about who owns
the allocation. That is solved by the inclusion of the TLLI in the PACKET UPLINK
ACK/NACK and PACKET DOWNLINK ACK/NACK messages. By doing so, the conten-
tion is resolved after the first occurrence of Packet Ack/Nack.

9.8.2.6 Uplink Access on PRACH (Access Persistence Control)


The mobile station makes at most MAX_RETRANS + 1 attempts to send a PACKET
CHANNEL REQUEST message. After sending each PACKET CHANNEL REQUEST
message, the mobile station listens to the full PCCCH that corresponds to the PRACH
(i.e., carried by the same PDCH).
The Control Parameters Information Element of PRACH contains the access persis-
tence control parameters, and it is broadcast on PBCCH and PCCCH. The parameters
included in the PRACH Control Parameters IE are:
– MAX_RETRANS, for each Radio Priority i (i=1,2,3,4); it defines (for each radio
priority) the maximum number of re-transmissions of the PACKET CHANNEL
REQUEST message; it corresponds to the GMANRETS database parameter
– PERSISTENCE_LEVEL, which consists of the PERSISTENCE_LEVEL P(i) for
each radio priority i (i = 1, 2, 3, 4), where P(i) is a value taken inside the {0, 1, …14,
16} group. If the PRACH Control Parameters IE does not contain the
PERSISTENCE_LEVEL parameter, this will be interpreted as if P(i)=0 for all radio
priorities. The user can set the four PERSISTENCE_LEVEL values (one for each
priority) by the following parameters: PERSTLVPRI1, PERSTLVPRI2,
PERSTLVPRI3 and PERSTLVPRI4
– S: corresponds to the GS database parameter
– TX_INT: corresponds to the GTXINT database parameter
The first attempt to send a PACKET CHANNEL REQUEST message, may be initiated
at the first possible TDMA frame containing PRACH, on the PDCH matching the mobile
station’s PCCCH_GROUP.

204 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

After each attempt, S and TX_INT parameters are used to determine the next TDMA
frame in which the MS is allowed to make a successive attempt. The number of TDMA
frames between two successive attempts to send a PACKET CHANNEL REQUEST
message, excluding the TDMA frames potentially containing the messages themselves,
is a random value drawn, for each transmission, with a uniform probability distribution in
the set {S, S + 1, …, S + TX_INT - 1}.
When the MS has made MAX_RETRANS + 1 attempts to send a PACKET CHANNEL
REQUEST message, the packet access procedure is aborted, a packet access failure
is indicated to upper layers and the mobile station returns to packet idle mode.
When the MS initiates a packet access procedure and receives from the network a
Packet Access Reject message from the network, corresponding to one of the 3 last
PACKET CHANNEL REQUESTs sent by the MS, it starts the T3172 timer; while the
timer is running, the MS is not allowed to access to the cell (i.e., it cannot send any other
PACKET CHANNEL REQUEST messages). Then one of the following situations can
occur (see Fig. 9.8):
– if the MS receives the Packet Uplink Assignment message, it stops the T3172 timer
– it the T3172 timer expires, the MS can start new transmissions of packet channel
requests

Start T3172 Stop T3172 T3172 Expired


(Reception of a Packet (Reception of a Packet Access in the cell is
Access Reject message) Uplink Assignment message) no longer prohibited

Fig. 9.8 Packet Access Reject Procedure

9.8.3 TBF Establishment Initiated by the Network on CCCH/PCCCH


When the GPRS/EGPRS MS is in standby state and the PCCCH channel is configured
in the serving cell, the mobile station listens to the paging sub-channels on PCCCH. If
PCCCH is not present in the considered cell, the mobile station listens to paging sub-
channels on CCCH.
Paging sub-channels are, in any case, monitored according to the paging groups
i defined for the mobile station and its current DRX mode (see "9.8.3.2 Discontinuous
Reception"). Paging for GPRS/EGPRS is performed on Routing Areas.

If the MS is in Standby state, the SGSN only knows the Routing Area on which the MS
is camped on. In order to initiate a downlink packet transfer, the SGSN sends to the MS
one or more PACKET PAGING REQUEST messages on the downlink PCH (PPCH).
The MS responds to one PACKET PAGING REQUEST message by initiating a mobile
originated packet transfer, as described in "9.8.2 TBF Establishment Initiated by the MS
on CCCH/PCCCH".
This mobile originated packet transfer allows the MS to send a PACKET PAGING
RESPONSE to the network. The packet paging response is one or more RLC/MAC data
blocks containing an arbitrary LLC frame.

A30808-X3247-L24-5-7618 205
GPRS/EGPRS Global Description Information
System

When the packet paging response has been sent by the MS and received by the
network, the mobility management state of the MS changes from standby to ready.
If the MS is already in READY state, the SGSN knows the exact cell where the MS is
i camped on; then the SGSN sends the assignment message on PCH (PPCH) or AGCH
(PAGCH), without sending the PACKET PAGING REQUEST message.
If an uplink packet transfer is in progress, the PACKET DOWNLINK ASSIGNMENT
message is transmitted on PACCH.

The transmission of RLC/MAC blocks to a MS in the ready state is initiated by the


network using the packet downlink assignment procedure. The network initiates the
packet downlink assignment procedure by sending the IMMEDIATE ASSIGNMENT
(PACKET DOWNLINK ASSIGNMENT) message on the CCCH (PCCCH) timeslot
corresponding to CCCH (PCCCH) group to which the mobile station belongs. If the
mobile station does not apply DRX, there is no further restriction on what part of the
downlink CCCH (PCCCH) timeslot an IMMEDIATE ASSIGNMENT (PACKET DOWN-
LINK ASSIGNMENT) message can be sent. If the mobile station applies DRX, the
message will be sent in a CCCH (PCCCH) block corresponding to a paging group deter-
mined for the mobile station in packet idle mode.
The downlink assignment message includes the list of PDCH(s) that will be used for
downlink transfer.
When the downlink TBF is set up, the following parameters and radio resources are allo-
cated to the MS:
– TFI
– Time Slot numbers
– ARFCN
– optionally Timing Advance parameters (TAI and Timeslot number)
– MAC access mode (always set to dynamic, see "9.8.1 Medium Access Modes")
– optionally, for EGPRS MSs, the EGPRS window size
The TBF establishment initiated by the network is shown in Fig. 9.9 in case of PCCCH.

206 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Fig. 9.9 TBF Establishment Initiated by the Network on PCCCH

9.8.3.1 Network Operation Modes for Paging


The network may provide co-ordination of paging for circuit-switched and packet-
switched services. Paging coordination means that the network sends paging messages
for circuit switched services on the same channel as used for packet switched services,
i.e.;
– on the PPCH paging channel, if the MS is in packet idle mode
– on the PDCH if the MS is in packet transfer mode
Then the MS must only monitor one channel at any time.
Three network operation modes are defined (see Tab. 9.2):
• Network operation mode I: the network sends CS paging messages for a
GPRS/EGPRS-attached MS, either on the same channel as the PS paging channel
(i.e., the PCCCH if it is configured, otherwise the CCCH), or on a PDCH traffic
channel. This means that the MS must monitor one paging channel, and that it
receives CS paging messages on the packet data channel when it has been
assigned a packet data channel. When the network operation mode I is used, the
network works in a coordinated way.
• Network operation mode II: the network sends CS paging messages for a
GPRS/EGPRS-attached MS on the CCCH paging channel, and this channel is also
used for PS paging. This means that the MS must monitor the CCCH paging
channel, but that CS paging continues on this paging channel even if the MS has
been assigned a packet data channel.
• Network operation mode III: the network sends CS paging messages for a
GPRS/EGPRS-attached MS on the CCCH paging channel, and sends PS paging

A30808-X3247-L24-5-7618 207
GPRS/EGPRS Global Description Information
System

messages either on PCCCH (if it is allocated in the cell) or on the CCCH paging
channel. Therefore a MS that wants to receive pages for both circuit-switched and
packet-switched services will monitor both paging channels if the packet paging
channel is allocated in the cell. No paging co-ordination is performed by the network.

Mode CS services paging PS paging channel Paging


channel co-ordination

Mode I PPCH paging channel PPCH paging channel YES


CCCH paging channel CCCH paging channel
Packet data channel Not Applicable
Mode II CCCH paging channel CCCH paging channel NO
Mode III CCCH paging channel PCCCH paging NO
channel
CCCH paging channel CCCH paging channel

Tab. 9.2 Network Operation Modes

When the Gs interface exists, all MSC-originated pages of GPRS/EGPRS-attached


MSs go via the SGSN, thus allowing network co-ordination of paging. Paging co-ordina-
tion is made by the SGSN based on the IMSI, and it is provided independently of
whether the MS is in STANDBY or READY state. The network operates in mode I.
When the Gs interface does not exist, all MSC-originated pages of GPRS/EGPRS-
attached MSs goes via the A interface, and co-ordination of paging cannot be
performed. Then the network will either:
– operate in mode II, meaning that the packet common control channel is not allocated
in the cell;
– operate in mode III, meaning that the packet common control channel is used for PS
paging when the packet paging channel is allocated in the cell.
The network operation modes (mode I, II, or III) are indicated as system information to
the MSs. The user sets this value with the NMO parameter. For proper operation, the
mode of operation should be the same in each cell of a routing area.
Based on the mode of operation provided by the network, the MS can then choose,
according to its capabilities, whether it can attach to PS services, to non-PS services, or
to both of them.

9.8.3.2 Discontinuous Reception


Paging is used to send paging information to mobile stations in packet idle mode (apart
from the current MM state, i.e., ready state or standby state), so a mobile station in
packet idle mode listens to radio blocks on CCCH or PCCCH.
For PS services, as for the CS service, it is also possible to organize paging channels
in combination with discontinuous reception (DRX). DRX allows MSs to reduce power
consumption.
A GPRS/EGPRS MS may be able to choose whether or not it wants to use discontin-
uous reception (DRX). If the MS uses DRX mode, the MS must also specify other DRX
parameters that indicate the delay for the network to send page requests or channel
assignments to the MS.

208 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

DRX parameters are indicated by the MS in the Attach procedure (see 9.3.2.1). Then
the then in each page request sends these parameters to the BSS, that uses both this
information and the IMSI of the MS to calculate the correct paging group.
In the GPRS attach procedure the following parameters are established:
• DRX/non-DRX indicator: indicates whether or not the MS uses DRX.
• DRX period: indicates the period of time between two consecutive paging blocks
(within a timeslot used as CCCH or PCCCH) where a MS, which is using DRX mode,
can receive its paging messages.
When PCCCH is used, the DRX period is defined by the SPLIT_PG_CYCLE param-
eter. The mobile station requests values for the SPLIT_PG_CYCLE parameter to be
applied on PCCCH.
The SPLIT_PG_CYCLE parameter handles the occurrence of paging blocks on
PCCCH monitored by the mobile station in DRX mode.
The support of the SPLIT_PG_CYCLE parameter on CCCH is optional. The
i SPGC_CCCH_SUP parameter (not configurable in the database) indicates the support
of the SPLIT_PG_CYCLE on CCCH from the network side.
If SPLIT_PG_CYCLE is not supported on CCCH, the period of monitoring paging blocks
on CCCHs is defined by the GSM NFRAMEPG parameter. The NFRAMEPG parameter
determines the number of 51 TDMA multiframes between two consecutive transmis-
sions of the same paging message in the same paging group.
The parameters used to define the paging groups for GPRS/EGPRS are shown in
Tab. 9.3, with the corresponding GSM parameters.

• Non-DRX timer: is used to determine the duration of the non-DRX mode period to
be applied by the mobile station when it has left the packet transfer mode and enters
the packet idle mode. A MS in non-DRX mode is required to monitor all of the radio
blocks of the PCCCH or CCCH channel; therefore the required time to execute the
paging procedure is reduced. As long as the timer is running (hence the MS is in
non-DRX mode), the BSC sends downlink assignments on the AGCH or PAGCH
(and not in paging blocks that the MS monitors when DRX mode is active, and that
occur with a low frequency) reducing the time to allocate resources.
So, when the MS changes from packet transfer mode to packet idle mode, the BSC
starts a timer; the duration of this timer is determined by the following:

timer = min (DRX_TIMER_MAX, NON_DRX_TIMER)

where:
– DRX_TIMER_MAX represents the DRXTMA parameter, and it is broadcast in the
cell
– NON_DRX_TIMER is a parameter requested by the MS in the PS attach proce-
dure
During this period, the MS is in non-DRX mode; when the timer expires, the MS
resumes discontinuous reception.

A30808-X3247-L24-5-7618 209
GPRS/EGPRS Global Description Information
System

In applications such as WAP, the time needed to get a reply is a key factor for end user
i acceptance. Because of highly interactive behavior of WAP with few seconds between
answers from network and subsequent user actions (for example navigating through
menus), response times are drastically reduced by sending the immediate assign-
ment/packet downlink assignment messages (polling requests) on the AGCH/PAGCH
instead of PCH/PPCH. So, when the DRX mode is temporarily disabled, the time that
occurs to receive at the MS side a data block that has been sent from the Gb interface,
is in average reduced by 50%.

When the mobile station receives a new value of the DRXTMA parameter, the
mobile station is not required to consider the new value until the next time it enters
packet idle mode.
There is another case when the MS enters the non-DRX mode: when initiating the MM
i procedures for PS attach and routing area update, the mobile station enters the MM
non-DRX mode period. This period ends when the corresponding MM procedures
terminate.

Subjects GPRS/EGPRS Corresponding GSM


parameters

PCCCH CCCH CCCH

DRX period SPLIT_PG_CYCLE NFRAMEPG (*) NFRAMEPG


SPLIT_PG_CYCLE (**)
Blocks not available for BSPBBLK + BPAGCHR NBLKACGR NBLKACGR (***)
paging per multiframe
Number of timeslots (phys- Depending on the number Depending on the number Depending on the
ical channels) containing of slots reserved for of slots reserved for number of slots reserved
paging PCCCH by means of the CCCH. for CCCH.
GDCH parameter.
(*) only when DRX period split is not supported.
(**) only when DRX period split is supported.
(***) NBLKACGR is a GSM parameter the indicates the number of CCCH blocks reserved for the access grant
signalling during a period of a 51 TDMA frames (that means during a period of a 51 TDMA multiframe). Its value
has to be always > 0, to support mobile stations which do not handle SI2quater correctly.

Tab. 9.3 Parameters for DRX Operation

GPRS/EGPRS Paging using CCCH


A mobile station using DRX is only required to monitor the PCH blocks belonging to its
paging group. The network sends the paging subchannel for a given MS every
NFRAMEPG multiframes, or every 64/SPLIT_PG_CYCLE multiframes if
SPLIT_PG_CYCLE is supported.
A mobile station not using DRX is required to monitor every PCH block on the same
CCCH as for DRX.
The network internal message flow is as follows:

210 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

1. the SGSN, which has the information about the usage of DRX, sends a paging
message to all PCUs that are supporting the right Routing Area. This message
includes the information on whether or not DRX is used, and additionally, if the
enhanced DRX mechanism is used, the SPLIT_PG_CYCLE parameter that indi-
cates that the existing DRX mechanism is supported by the network;
2. the PCU forwards the Packet Paging Request message combined with the
requested paging parameters over the internal interface to the TDPC;
3. the TDPC calculates the right paging group and forwards per LAPD connection the
Packet Paging Request message to the paging queues inside the BTS. Additionally
the BSC evaluates all needed DRX parameters that must be broadcast on the
BCCH;
4. the BTS queues all Packet Paging Request messages and sends them in first-in
first-out order on the PCHs in the CCCH multiframe.

GPRS/EGPRS Paging using PCCCH


A MS using DRX is required to monitor the PPCH. The network sends the paging
subchannel for a given MS every 64/SPLIT_PG_CYCLE multiframes.
A mobile station not using DRX is required to monitor every PPCH block on the same
PCCCH as for DRX.
The network internal message flow as follows:
1. The SGSN, which has the information about the usage of DRX, sends a paging
message to all PCUs located in the right Routing Area. This message includes the
information about whether or not DRX is used and additionally (if the enhanced DRX
mechanism is used) the SPLIT_PG_CYCLE parameter;
2. The PCU calculates the right paging group and adds all Packet Paging Request
messages on its paging group queues. Additionally, the PCU evaluates all needed
DRX parameters that must be broadcast on PBCCH;
3. The PCU includes the Packet Paging Request messages into RLC/MAC blocks and
schedules the messages into the PDCH multiframes, which contain PCCCH. The
RLC/MAC blocks are transferred via PCU frames to the BTS, which immediatly
transmits the Packet Request message.

9.8.4 Relative Reserved Block Period Field (RRBP)


The RRBP field is contained in the MAC header of every RLC/MAC block sent in down-
link direction. Its value specifies a single uplink block in which the mobile station will
transmit either a PACKET CONTROL ACKNOWLEDGEMENT message (to acknowl-
edge a downlink control block) or another PACCH block to the network.
If the RRBP field is received as part of a RLC/MAC control block containing any
message except Packet Paging Request, Packet Access Reject, and Packet Queueing
Notification, the mobile station will transmit a PACKET CONTROL ACKNOWLEDGE-
MENT message in the specified uplink radio block.
If the RRBP field is received as part of a RLC/MAC control block containing a Packet
Paging Request, Packet Access Reject, or Packet Queueing Notification message, the
mobile station ignores this RRBP field.
If the RRBP field is received as part of a RLC/MAC data block, the mobile station will
transmit a PACCH block (e.g., a PACKET UPLINK ACK/NACK message to acknowl-
edge the downlink data block) in the specified uplink radio block.

A30808-X3247-L24-5-7618 211
GPRS/EGPRS Global Description Information
System

The mobile station will always transmit the uplink radio block on the same timeslot as
the block where the RRBP has been received.
To indicate to the MS whether or not the received RRBP field is valid, a bit of the MAC
header is used; according to the value of this bit, the MS knows if, in the received block,
the RRBP field is meaningful.
The multislot class of the MS limits allowed combinations and configurations when the
i MS supports multislot communications. When an MS has established a downlink TBF,
it cannot transmit in uplink direction (after a polling by the network) on any timeslot; in
fact for each mobile station, according to its multislot class, downlink and uplink
timeslots usage is specified (see "4.7.1 Mobile Station Classes for Multislot Capabili-
ties"). Therefore, to poll the MS, the network must send the downlink block with a valid
RRBP field on a timeslot where the polled MS is able to answer.

9.8.5 Polling Procedures


As described in "4.6 Packet Timing Advance Estimation", the initial timing advance esti-
mation is based on the single access burst carrying the Packet Channel Request. The
Packet Uplink Assignment or Packet Downlink Assignment (or the Immediate Assign-
ment if the PCCCH is not configured) carries the estimated timing advance value to the
MS.
But, when Packet Downlink Assignment must be sent without prior paging (i.e., in the
Ready state), no valid timing advance value may be available.
Then the network has two options:
1. The RRBP field of the Packet Downlink Assignment message can be set to trigger
the transmission of the PACKET CONTROL ACKNOWLEDGEMENT message (see
"9.8.4 Relative Reserved Block Period Field (RRBP)"). This can be used only if the
System information indicates that acknowledgement is access bursts, i.e., it can be
used only if CACKTYP is set to 0;
The CACKTYP parameter is hard-coded to the value 0.
i
2. Packet Downlink Assignments can be sent without timing advance information. In
that case, it is indicated to the MS that it can only start the uplink transmission after
the timing advance value has been obtained by the continuous timing advance
update procedure.
The continuous timing advance update procedure can create some delays between the
packet downlink assignment message and the beginning of data transfer in downlink
direction (this delay is due to the time needed to exchange timing advance information
between the network and the MS). In order to reduce the time between a packet down-
link assignment message and the effective beginning of downlink data transmission, a
polling procedure is executed by the network to order the MS to send a Packet Control
Acknowledgment message, formatted as four Access Bursts. This procedure foresees
polling the MS by means of the RRBP field of the assignment message; as a conse-
quence, the CACKTYP parameter must be set to 0, to force the MS to respond with
four Access Burst.
Different procedures are executed according to whether or not the PBCCH is config-
ured.

212 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Procedure when PBCCH is configured


If PBCCH is configured, the following steps are executed:
– the BSC sends the PACKET DOWNLINK ASSIGNMENT message (on the
PPCH/PAGCH) setting the RRPB field to poll the MS;
– upon reception of the PACKET CONTROL ACKNOWLEDGEMENT message from
the MS, the BSC can immediately start downlink transmission of data blocks;
– as soon as possible (also before the transmission of the first downlink data block)
the BSC must send the PACKET POWER CONTROL/TIMING ADVANCE message
to the MS, including the estimated timing advance value.

Procedure when PBCCH is not configured


If PBCCH is not configured, the following steps are executed:
– the BSC sends the IMMEDIATE ASSIGNMENT message for downlink TBF on
PCH/AGCH;
– then the BSC sends on the assigned PDCH(s) a PACKET DOWNLINK ASSIGN-
MENT message setting the RRPB field to poll the MS;
– upon reception of the PACKET CONTROL ACKNOWLEDGEMENT message from
the MS, the BSC can immediately start downlink transmission of data blocks;
– as soon as possible (also before the transmission of the first downlink data block)
the BSC must send the PACKET POWER CONTROL/TIMING ADVANCE message,
including the estimated timing advance value, to the MS.

9.9 RLC Data Block Transfer


The RLC functions support two modes of operation:
• RLC acknowledged mode: this mode is used for data applications where the
payload content must be preserved. It is the typical mode for Background class
(background delivery of e-mails, SMS, download of databases) and Interactive class
applications (web browsing).
• RLC unacknowledged mode: this mode is used for delay-sensitive services, such as
Conversational class (voice, video conference) and Streaming class applications
(one-way real time audio and video).
A TBF may operate in either RLC acknowledged mode or RLC unacknowledged mode.

9.9.1 Acknowledged Mode for RLC/MAC Operation


There are some differences regarding GPRS and EGPRS acknowledged modes; these
differences are described below.

9.9.1.1 GPRS Acknowledged Mode


GPRS acknowledged operation mode uses retransmission of RLC data blocks to
achieve high reliability. The transfer of RLC Data Blocks in the RLC/MAC acknowledged
mode is controlled by a selective ARQ mechanism (type I ARQ) coupled with the
numbering of the RLC Data Blocks within one Temporary Block Flow. The sending side
(the MS or the network) transmits blocks within a window (the length of the window is
fixed to 64 blocks) and the receiving side sends a Packet Uplink Ack/Nack or a Packet
Downlink Ack/Nack message when needed. The transmitting side numbers the RLC
data blocks via the block sequence number (BSN), which is used for retransmission and
reassembly. Every such message acknowledges all correctly received RLC Data Blocks

A30808-X3247-L24-5-7618 213
GPRS/EGPRS Global Description Information
System

up to an indicated block sequence number (BSN), thus “moving” the beginning of the
sending window on the sending side. Additionally, the bitmap that starts at the same
RLC Data Block is used to selectively request erroneously received RLC Data Blocks
for retransmission. The sending side then retransmits the erroneous RLC Data Blocks,
eventually resulting in further sliding of the sending window. A missing Packet Ack/Nack
is not critical and a new one can be issued whenever.

9.9.1.2 EGPRS Acknowledged Mode


The transfer of RLC Data Blocks in the EGPRS acknowledged RLC/MAC mode can be
controlled by the following:
a) a selective type I ARQ mechanism, where coding of a RLC Data Block is solely
based on the prevailing transmission (i.e., erroneous blocks are not stored);
b) a type II hybrid ARQ mechanism (called Incremental Redundancy - IR) where erro-
neous blocks are stored by the receiver and a joint decoding with new transmissions
is done.
Both methods are coupled with the numbering of the RLC Data Blocks within one
Temporary Block Flow.
In Release BR7.0 , the Incremental Redundancy mechanism for EGPRS is only used in
i the downlink direction; in the uplink direction, only the selective type I ARQ mechanism
is used.

The sending side (the MS or the network) transmits blocks within a window and the
receiving side sends Packet Uplink Ack/Nack or Packet Downlink Ack/Nack message
when needed.
For EGPRS, the window size (WS) will be set by the operator according to the number
of timeslots allocated in the direction of the TBF (uplink or downlink). The operator can
set the window sizes with the following parameters:
– EGWSONETS, in case one timeslot is assigned
– EGWSTWOTS, in case two timeslots are assigned
– EGWSTHREETS, in case three timeslots are assigned
– EGWSFOURTS, in case four timeslots are assigned
– EGWSFIVETS, in case five timeslots are assigned
– EGWSSIXTS, in case six timeslots are assigned
– EGWSSEVENTS, in case seven timeslots are assigned
– EGWSEIGHTTS, in case eight timeslots are assigned
According to the link quality, an initial MCS is selected for an RLC block. For retransmis-
sion, the same or another MCS from the same family of MCSs can be selected (see
"10.5 Link Adaptation"). For example if MCS-7 is selected for the first transmission of an
RLC block, any MCS of the family B can be used for retransmissions.
RLC data blocks initially transmitted with MCS4/MCS5 or MCS6/MCS7/MCS8 or MCS9,
can optionally be retransmitted with MCS1, MCS2, and MCS3 respectively, using two
radio blocks. In this case, the split block field in the header is set to indicate that the RLC
data block is split, and the order of the two parts. For blocks initially transmitted with
MCS8 that are retransmitted using MCS6 or MCS3, padding of the first six octets in the
data field will be applied, and the CPS field will be set to indicate that this has been done.
Incremental redundancy is used only in the downlink direction. The split block field is
used to indicate to the MS whether or not the block has been segmented. In fact, the
following must be noted:

214 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

a) when ARQ mode type I is used, the retransmission is executed with a coding
scheme of the same family of the block received with errors, and block splitting is
possible;
b) when ARQ mode type II is used, the retransmission is executed with a coding
scheme of the same family of the block received with errors (but with another punc-
turing scheme) and block splitting is not allowed;
In the EGPRS type II Hybrid ARQ mode, the information is first sent with one of the initial
code rates (i.e., the rate 1/3 encoded data is punctured with the puncturing scheme (PS)
1 of the selected MCS). If the RLC Data Block is received in error, additional coded bit
(i.e., the output of the rate 1/3 encoded data that is punctured with PS 2 of the prevailing
MCS) are sent and decoded with the previously received code-words until decoding
succeeds. If all of the code-words (different punctured versions of the encoded data
block) have been sent, the procedure will start over, and the first code-word (which is
punctured with PS 1) will be sent followed by PS 2 etc.
RLC data blocks, that are retransmitted using a new MCS, will be sent with the punc-
turing scheme indicated in Tab. 9.4, at the first transmission after the MCS switch.

MCS switched MCS switched to PS of last transmis- PS of first trans-


from sion before MCS mission after MCS
switch switch

MCS9 MCS6 PS1 or PS3 PS1


PS2 PS2
MCS6 MCS9 PS1 PS3
PS2 PS2
MCS7 MCS5 any PS1
MCS5 MCS97 any PS2
all other combinations any PS1

Tab. 9.4 Puncturing Schemes to be used after a Coding Scheme Switch

In the EGPRS type I ARQ, the operation is similar to one of the EGPRS type II hybrid
ARQ, except that the decoding of an RLC Data Block is solely based on the prevailing
transmission (i.e., erroneous blocks are not stored).
Therefore, the MS can use either the type I ARQ or the type II ARQ mode, according to
the current situation .
If the memory for IR operation run out in the MS, the MSwill indicate this by setting the
LA/IR bit in the EGPRS PACKET DOWNLINK ACK/NACK message.
If IR is considered as "not-working-properly"at the MS (IR_statusk<0.5, see "10.5 Link
Adaptation"), then the PCU may decide to re-segment the not acknowledged blocks.
Therefore, for retransmissions, an MCS within the same family as the initial transmission
may be used and the payload may be split. On the contrary, if IR is considered as "prop-
erly working" (IR_statusk>0.5) at the MS, retransmissions may be realized with an MCS
within the same family as the initial transmission without splitting the payload.
Furthermore, it is mandatory for an EGPRS MS receiver to be able to perform joint
decoding among blocks with different MCSs if the combination of MCSs is one of the
following:

A30808-X3247-L24-5-7618 215
GPRS/EGPRS Global Description Information
System

– MCS5 and MCS7


– MCS6 and MCS9

9.9.2 Unacknowledged Mode for RLC/MAC Operation


RLC unacknowledged operation mode does not include any retransmission. The
transfer of RLC Data Blocks in the RLC/MAC unacknowledged mode is controlled by the
numbering of the RLC Data Blocks within one Temporary Block Flow. The block
sequence number (BSN) in the RLC data block header is used to number the RLC data
blocks for reassembly. The receiving side extracts user data from the received RLC
Data Blocks and attempts to preserve the user information length by replacing missing
RLC Data Blocks by dummy information bit. To convey the necessary control signalling
(e.g., monitoring of channel quality for downlink channel or timing advance correction for
uplink transfers) temporary acknowledgement messages are transmitted, with the same
mechanisms and the same message format used by the RLC/MAC acknowledged
mode. The fields for denoting the erroneous RLC blocks may be used as an additional
measure for channel quality (i.e., parameter for link adaptation). The sending side (the
MS or the network) transmits a number of radio blocks and then polls the receiving side
to send an acknowledgement message. A missing acknowledgement message is not
critical and a new one can be obtained whenever.
When working in RLC unacknowledged mode, badly received blocks are not re-
transmitted (ARQ functions are not used). BLER information could be derived and the
link adaptation algorithm could work in a similar way to the acknowledged mode case.
However the unacknowledged mode is typically used to support real time applications
and in this case minimizing the data unit error ratio, i.e., the fraction of data unit lost or
detected as erroneous, is much more important than maximizing throughput.

9.9.3 Operations on Uplink TBF

9.9.3.1 Uplink TBF Using the Acknowledged Mode


The mobile station transmits RLC/MAC blocks in each assigned uplink data block.
RLC/MAC control blocks take precedence over RLC data blocks, i.e., temporarily
replacing the PDTCH with PACCH. The network sends PACKET UPLINK ACK/NACK
messages when needed.
In case of GPRS service, where the window size is fixed to 64 blocks (see
"9.9.1.1 GPRS Acknowledged Mode"), to indicate to the network when it must send a
PACKET UPLINK ACK/NACK, the NRLCMAX parameter is defined: by this parameter,
the user can configure how many blocks must be transmitted by the MS, before
receiving a PACKET UPLINK ACK/NACK message. The NRLCMAX value is chosen as
a compromise between two necessities:
– it will avoid reaching the stall condition of the transmitting window (see below); from
this point of view the value should be quite low;
– it will avoid a frequent number of PACKET UPLINK ACK/NACK messages; from this
point of view, the value should be quite high.
In case of EGPRS service, where, according to the number of timeslots assigned to the
MSs, a specific window size is used (see "9.9.1.2 EGPRS Acknowledged Mode"), to
indicate to the network when it must send a PACKET UPLINK ACK/NACK, different
parameters are provided:

216 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

– EGPLGPONETS, in case one timeslot is assigned;


– EGPLGPTWOTS, in case two timeslots are assigned;
– EGPLGPTHREETS, in case three timeslots are assigned;
– EGPLGPFOURTS, in case four timeslots are assigned;
– EGPLGPFIVETS, in case five timeslots are assigned;
– EGPLGPSIXTS, in case six timeslots are assigned;
– EGPLGPSEVENTS, in case seven timeslots are assigned;
– EGPLGPEIGHTTS, in case eight timeslots are assigned.
By these parameter the user can configure how many blocks must be transmitted by the
MS, before receiving a PACKET UPLINK ACK/NACK message, according to the
number of assigned PDCHs. The value of these parameters is chosen as a compromise
between two necessities:
– avoid reaching the stall condition of the transmitting window;
– avoid a frequent number of PACKET UPLINK ACK/NACK messages.
If the mobile station does not receive PACKET UPLINK ACK/NACK messages that
allows it to advance the transmitting window, the transmitting window stall condition is
reached: upon detecting this condition, the mobile station sets the Stall indicator (SI) bit
in all subsequent uplink RLC data block, until the stall condition ceases to exist (i.e.,
when a valid PACKET UPLINK ACK/NACK message is received).
Upon detecting the stall condition, the mobile station also starts the T3182 timer. T3182
timer is stopped upon the reception of a PACKET UPLINK ACK/NACK message, that
allows terminating the stall condition (see Fig. 9.10).
If T3182 timer expires:
– the mobile station decrements the N3102 counter by the PAN_DEC value; the
PAN_DEC value is defined by the PKTNDEC parameter;
– the mobile station aborts all TBFs in progress and its associated resources;
– if N3102 counter has not reached the value 0, the mobile station returns to the
CCCH or PCCCH and initiates the establishment of a new uplink TBF, otherwise the
MS performs an abnormal release with cell reselection (see below).
Whenever the mobile station receives a PACKET UPLINK ACK/NACK message that
allows the advancement of the sending window, the mobile station increments the
N3102 counter by the PAN_INC value, however N3102 will never exceed the PAN_MAX
value.
The user can configure PAN_INC and PAN_MAX values by PKTNINC and PKTNMA
parameters respectively.
Upon cell reselection, the mobile station sets the N3102 counter to the PAN_MAX value.
When N3102 = 0 is reached, the mobile station performs an abnormal release with cell
re-selection (see Abnormal Cell Re-selection).
If PAN_DEC, PAN_INC, or PAN_MAX are set to the value 0, counter N3102 is disabled.
i

A30808-X3247-L24-5-7618 217
GPRS/EGPRS Global Description Information
System

N3102 = N3102 + PAN_INC N3102 = N3102 - PAN_DEC

Cell reselection Start T3182 Stop T3182 T3182 Expired


N3102 = PAN_MAX Stall condition (Reception of a Packet Abort all TBFs
Uplink Acknowledge)

Fig. 9.10 Behavior of T3182 Timer and N3102 Counter

9.9.3.2 Uplink TBF Using the Unacknowledged Mode


When the unacknowledged mode is used, the network sends PACKET UPLINK
ACK/NACK messages when needed, while the mobile station sets the Stall indicator bit
to ‘0’ in all RLC data blocks. In unacknowledged mode, the number of blocks to be trans-
mitted by the MS, before receiving a control message from the network, is fixed to 64
(i.e., the value of the transmitting window used in acknowledged mode).
If the mobile station transmits a number of data blocks equal to the window size, without
receiving a PACKET UPLINK ACK/NACK message, the mobile station starts T3182
timer. T3182 will be stopped upon reception of a PACKET UPLINK ACK/NACK
message.
If timer T3182 expires, the mobile station decrements the N3102 counter by the
PAN_DEC (PKTNDEC) value, aborts all TBFs in progress and its associated resources,
returns to CCCH or PCCCH and initiates the establishment of a new uplink TBF.
Whenever the mobile station receives a PACKET UPLINK ACK/NACK message, the
mobile station will increment N3102 by PAN_INC (PKTNINC), however N3102 will never
exceed the value PAN_MAX (PKTNMA).
Upon cell reselection, the mobile station will set counter N3102 to the PAN_MAX value.
When N3102 = 0 is reached, the mobile station performs an abnormal release with cell
re-selection.
If PAN_DEC, PAN_INC, or PAN_MAX are set to the value 0, the N3102 counter is
i disabled.

9.9.3.3 Anomalies During an Uplink TBF


Mobile Station Side
When the mobile station transmits an RLC/MAC block to the network, it starts timer
T3180. When the mobile station detects an assigned USF value on an assigned PDCH,
the mobile station reset the T3180 timer. If the T3180 timer expires, the mobile station
aborts all TBFs in progress and its associated resources, returns to the CCCH or
PCCCH, and initiates the establishment of a new uplink TBF.

218 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Network Side
Whenever the network receives a valid RLC/MAC block from the mobile station, it resets
the N3101 counter. The network increments the N3101 counter for each allocated radio
block to that mobile station, for which no data is received.
If N3101 reaches the N3101 maximum value, the network stops the scheduling of
RLC/MAC blocks from the mobile station and starts the T3169 timer. When T3169
expires, the network may reuse the USF and TFI values (the procedure is shown in
Fig. 9.11).
The user can also define the N3101 maximum value by the N3101 parameter.

N3101 = 0 N3101 = N3101 + 1 N3101 = N3101+1= N3101max

time

NW sends Data is NW sends Data is not NW sends USF Data is not received T3169 Expired
USF received USF received from the MS. Reuse of TFI
from the from the Start T3169 and USF
MS. MS. communication with
MS is broken.

Fig. 9.11 Detection of Anomalies during an Uplink TBF on the Network Side

9.9.3.4 Release of an Uplink TBF


The release of resources is normally initiated from the MS by counting down the last
blocks. For the normal release of resources of a RLC connection, carrying a mobile orig-
inated packet transfer, the mechanism based on the acknowledgment of the final Packet
Uplink Ack/Nack message (combined with timers) is used (see Fig. 9.12).
The MS initiates the release of the uplink TBF by beginning the countdown process. The
MS sends the Countdown Value (CV) in each uplink RLC data block to indicate to the
network the absolute BSN (Block Sequence Number) of the last RLC data block, that
will be sent in the uplink TBF.
The CV value is calculated as follows:

x = round  ----------------------------------------
TBC – BSN′ – 1
 NTS 

then

x if x ≤ BSCVMAX
CV = 
 15 otherwise

A30808-X3247-L24-5-7618 219
GPRS/EGPRS Global Description Information
System

where:

- TBC = total number of RLC data blocks that will be transmitted in the TBF;
- BSN’ = absolute block sequence number of the RLC data block, with range from 0
to (TBC - 1);
- NTS = number of timeslots assigned to the uplink TBF, with range 1 to 8;
- BSCVMAX is the BSCDVMA parameter, broadcasted in the system information;
- the round() function rounds upwards to the nearest integer; the division operation
is non-integer and the result is 0 only for (TBC - BSN’ - 1) = 0.

The final RLC data block transmitted in the TBF (i.e., the RLC data block with BSN’ =
TBC - 1) will have CV set to the value ‘0’. Once the mobile station transmits a value of
CV other than 15, the MS will not queue any new RLC data blocks, and any data that
arrives after the commencement of the countdown process will be sent within a future
TBF.
After the MS has sent its last RLC Data Block (indicated by the countdown field), the
acknowledgement is expected from the network side. By sending the last block, the MS
may no longer use the same assignment, unless a negative acknowledgement arrives.
It also means that the network side may reallocate the same USF(s) to another user as
soon as all of the RLC Data Blocks belonging to that Temporary Block Flow are correctly
received. When sending the last RLC data block, the MS starts also the T3182 timer.
Then the network, if all RLC Data Blocks have been correctly received, sends the Packet
Uplink Ack/Nack message to the MS that must be immediately acknowledged by the MS
in the reserved uplink block period (the network also resets the N3103 counter).
If T3182 timer expires, before the MS receives the Packet Uplink Ack/Nack message,
then the mobile station aborts all TBFs in progress and its associated resources, returns
to the CCCH/PCCCH and initiates the establishment of a new uplink TBF.
When the MS receives the Packet Uplink Ack/Nack message, it responds to the network
by the Packet Control Acknowledgment message in the reserved uplink block period.
Upon reception of the acknowledgement, the network can reuse the TFI and USF
values.

220 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Fig. 9.12 Release of an Uplink TBF

If the network does not receive the PACKET CONTROL ACKNOWLEDGEMENT


message in the radio block indicated by the RRBP field (see "9.8.4 Relative Reserved
Block Period Field (RRBP)"), it increments the N3103 counter and retransmits the
PACKET UPLINK ACK/NACK message. If counter N3103 exceeds its limit, which the
user can define with the N3103 parameter, the network starts the T3169 timer. When
the T3169 timer expires, the network may reuse the TFI and USF resources
(see Fig. 9.13).

N3103 = N3103 + 1 N3103 = N3103+1= N3103max

NW waits for NW has not received any NW has not received any T3169 Expired
acknowledgment ackn. from the MS. ackn. from the MS. Reuse of TFI and USF
Start T3169
communication with
MS is broken.

Fig. 9.13 Release of Resources on the Network Side during an Uplink TBF (in case of T3169 timer expiration)

A30808-X3247-L24-5-7618 221
GPRS/EGPRS Global Description Information
System

Improving overall performances in the interaction between TCP/IP based applica-


tions and the GPRS/EGPRS network (uplink direction)
In the PS network, the continuous interruptions of data flow due to frequent establish-
ments and releases of TBFs, reduce the performances of many TCP/IP based applica-
tions, such as Web Browsing applications.
To reduce these disadvantages during uplink data transfer, a delay into the uplink TBF
release procedure is introduced (see Fig. 9.12).
The BSC delays sending the final PACKET UPLINK ACK/NACK message. This
behavior is introduced to save time when a downlink LLC PDU arrives from SGSN just
after the reception of the RLC Block with CV = 0. In this case, the BSC can open the
downlink TBF as a concurrent case, that is faster than the normal procedure (where
normal means assignment messages sent on control channel, after the PACKET
CONTROL ACKNOWLEDGEMENT message that has closed the uplink TBF).
This delay is a fixed value (400 msec.), and it is not longer because during this delay,
the MS cannot require the opening of a new uplink TBF.
If during the delay time a new downlink TBF is opened, final PACKET UPLINK
ACK/NACK is sent without waiting for the expiration of delay timer.

9.9.4 Operations on Downlink TBF

9.9.4.1 Acknowledged and Unacknowledged Modes on Downlink TBFs


The mobile station receives RLC/MAC blocks on the assigned downlink PDCHs. On
each assigned PDCH, the mobile station decodes the TFI value, and decodes the RLC
data blocks intended for the mobile station.
In acknowledged mode, to indicate to the MS when it must send a PACKET DOWNLINK
ACK/NACK message different parameters are provided according to the service type:
• in case of GPRS, the NRLCMAX parameter is provided; by this parameter, the user
can configure how many blocks must be transmitted by the network, before
receiving a PACKET DOWNLINK ACK/NACK message;
• in case of EGPRS eight parameters are provided (depending on the defined window
size (see 9.9.1.2):
– EGPLGPONETS, in case of one timeslot assigned;
– EGPLGPTWOTS, in case of two timeslots assigned;
– EGPLGPTHREETS, in case of three timeslots assigned;
– EGPLGPFOURTS, in case of four timeslots assigned;
– EGPLGPFIVETS, in case of five timeslots assigned;
– EGPLGPSIXTS, in case of six timeslots assigned;
– EGPLGPSEVENTS, in case of seven timeslots assigned;
– EGPLGPEIGHTTS, in case of eight timeslots assigned;
by these parameters the user can configure how many blocks have to be transmitted
by the network, before receiving a PACKET DOWNLINK ACK/NACK message,
according to the number of PDCHs assigned to the MS.
In unacknowledged mode, the MS must send a PACKET DOWNLINK ACK/NACK
message after:
• it has received 64 blocks (i.e., the window size) in case of GPRS;
• it has received a number of blocks equal to the configured window size, in case of
EGPRS.

222 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

In both unacknowledged cases (GPRS and EGPRS) the PACKET DOWNLINK


ACK/NACK message is used to check the connection between the MS and the network
(see below).
For both of the operation modes, a control procedure is used by the network to verify if
the MS is correctly receiving the downlink RLC/MAC blocks. The network, using the
RRBP field (see "9.8.4 Relative Reserved Block Period Field (RRBP)"), reserves the
uplink resource to transmit control messages.
The N3105 parameter implements the threshold for unreceived control messages from
the MS, after sending the RRPB field in downlink direction. If the threshold is reached,
the communication with the associated MS is broken.
Every time the network does not receive the control message from the MS, the N3105
counter is increased; every time the network receives the control message from the MS,
the N3105 counter is reset. When the N3105 counter reaches its maximum value
(N3105max), the communication with the MS is broken (see Fig. 9.14).
The user can configure the N3105max threshold with the N3105 parameter.

N3105 = N3105 + 1 Reset N3105 N3105 = N3105 +1= N3105max

NW sets RRPB NW has not received any NW has received a NW has not received any
in DL data block control message control message control message
from the MS. from the MS. from the MS.
Communication with
MS is broken.

Fig. 9.14 Control Procedure Executed by the Network during a Downlink TBF

9.9.4.2 Release of a Downlink TBF


The release of resources (see Fig. 9.15) is initiated by the network by terminating the
downlink transfer and polling the MS for a final Packet Downlink Ack/Nack message.
The network indicates the last downlink RLC data block by setting the Final Bit Indicator
bit (FBI) to 1.
It is possible for the network to change the current downlink assignment by using the
Packet Downlink Assignment or Packet Timeslot Reconfigure message, which then
must be acknowledged by the MS in a reserved radio block on the uplink.
The TFI handling is steered with timers that run on both the MS and the network sides:
– after having sent the last RLC Data Block to the MS, the network starts the T3191
timer
– when the MS receives the last RLC Data Block, the MS starts the T3192 timer; when
it expires, the current assignment becomes invalid for the MS
Therefore, upon reception of the final Packet Downlink Ack/Nack from the MS (with Final
Ackn = 1), the T3193 timer is started on the network side (and the T3191 timer is

A30808-X3247-L24-5-7618 223
GPRS/EGPRS Global Description Information
System

stopped). When it expires, the current assignment becomes invalid for the network, and
TFI can be reused by the network.

Fig. 9.15 Release of a Downlink TBF

If the mobile station (in acknowledged mode), after having received a RLC data block
with FBI=1, transmits a PACKET DOWNLINK ACK/NACK message with the Final Ack
indicator not set to 1, it will continue to monitor all assigned PDCHs; in this case the
network must retransmit some RLC blocks.
If the network receives a PACKET DOWNLINK ACK/NACK message before the T3191
timer expires, and if retransmissions are required, then the network stops the T3191
timer and retransmits necessary RLC data blocks.

Improving overall performances in the interaction between TCP/IP based applica-


tions and the GPRS/EGPRS network (downlink direction)
As described in 9.9.3.4, the continuous interruptions of data flow due to frequent estab-
lishments and releases of TBFs, reduce the performances of many TCP/IP based appli-
cations.
To reduce these disadvantages during downlink data transfer, a delay into the downlink
TBF release procedure is introduced (see Fig. 9.15).
The BSC, before sending the FBI=1, waits for the expiration of a timer. The operator can
set this timer using the TIMEDTBFREL attribute.
Setting the TIMEDTBFREL attribute to 0 means no delay in TBF downlink releases.
i
During this time, the BSC maintains opened the downlink TBF, by sending Dummy LLC
frames included in single RLC Blocks, with the polling bit (RRBP) set to one.
There are two advantages when Dummy LLC frames are used:

224 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

1. The MS may request UL resource in the PACKET DOWNLINK ACK/NACK


message; with these Dummy frames, the BSC presses the MS for an answer, i.e.,
if the MS has data to send, it will send it quickly.
The delay time is not only influenced by the network, but also by the MS, so the delay
time can be different using different mobile stations;
2. If a new LLC should arrive from the SGSN, it is possible to send the relative RLC
block immediately (without an explicit assignment procedure). In this manner there
is an unique long single DL TBF, rather than several DL TBFs with a lot of assign-
ment messages.
In order to speed up the uplink establishment during the Delay TBF Release time, polling
periods should be quite low.
In current release, the delay between two subsequent pings is reduced due to the fact
that the block containing the Dummy LLC is sent (polled) every 50 ms (alternating 40/60
ms) for a duration of 450 ms; in this way MS may request resources in a shorter time.
After that the polling period is extended to 280 ms (1 poll every 280 ms). The reason is
a compromise between speed up of procedure, battery saving and resource usage. With
this strategy the MS uplink establishment is speed up during the delay downlink TBF
release time; infact improvements of about 80 ms are expected on Ping Delay time
measured as a sequence of 50 pings) and also on FTP throughput.
The Ping Delay time is further reduced, by reducing the internal PCU queue of radio
blocks, obtaining a quite immediate sending of data when available. This improvement
can save approximately 20-40 msec per direction and it is implemented only on PPXUs.
A further improvement for the current release is related to the First Ping. The worst
condition in a cell with all the PDCHs being idle is a a value of 1.55 seconds for the First
Ping. The implemented solution is general and it applies also to other GMM and SM
procedures as well as to Data Transfer. The number of PDT assigned to a single block
has to be set to the value “2” if concatenated PCU frames are used in the cell. In case
standard PCU frames are used in the cell the number of PDT has to be set to the value
“1”. Furthermore also the RLC octet count in PRR has to be taken into account for
avoiding an unnecessary alignment of 2 Uplink Timeslot (for Mobile Stations that have
more than 1 Timeslot in Uplink direction).
With this improvement two sequential PDT/PDCH alignment are executed instead of
three. The improvement of the First Ping is expected to be very near to 200 msec and it
is valid for each GPRS and also for almost all EGPRS access cases.

9.9.5 Notes About Concurrent TBFs


When concurrent TBFs must be established during either the resource allocation or the
resource upgrade strategy (see "5.3 Management of Packet Data Channels"), and when
the multislot class of the mobile station allows a degree of freedom on how to assign
resources between uplink and downlink, it is necessary to manage resources in an
optimal way to balance the traffic of the two directions.
Mobile stations for which this problem arises are those providing a dynamic allocation
i of the number of resources, i.e., those belonging to the following multislot classes
(see Tab. 4.6): 6, 7, 10, 11, 12.
For example class 8 (4+1) is not affected, because in total 5 "timeslots" can operate;
instead class 10 (4+2) is affected, because in sum only 5 timeslots can be active,
however 6 are possible. So either 4+1 or 3+2 operation (or less) is possible.

A30808-X3247-L24-5-7618 225
GPRS/EGPRS Global Description Information
System

Tests with a mobile station with multislot class 6 have shown that with two simultaneous
FTP connections, one in uplink and the other in downlink direction (duplex FTP), in case
of downlink preferred configuration (3+1) the downlink throughput is worse than in uplink
preferred configuration (2+2). This is due to the fact that FTP connections are based on
the TCP transfer protocol, which causes acknowledged traffic in the opposite direction.
Because of the delayed acknowledgement packets (caused by the queue in MS or note-
book, which is always full concerning the uplink traffic) the downlink transfer is reduced
(stalled condition).
If the chosen solution was always downlink biased (i.e., 3+1), also pure uplink traffic
such FTP put would not be handled optimally, since the network would change to down-
link preferred allocation as soon as first downlink TBFs for TCP/IP acknowledgments
arrives.
The current implementation to manage concurrent TBFs is as follows.
When a downlink data transfer is set up, data transfer is always allocated with downlink
priority.
Regarding the uplink direction, the mobile station might request the uplink TBF with
either the Packet_Resource_Request or the Packet_Downlink_Ack/Nack message
(PDAN). Within these messages there is the Channel_Request_Description information
element that contains a field called RLC_Octet_Count.
The RLC_Octet_Count field indicates the number of RLC data octets that the mobile
station wishes to transfer; its range is from 0 to 65535, and:
– the value 0 is interpreted as a request for an open-ended TBF by the mobile station
(i.e., the mobile station does not specify the number of blocks it must transmit);
– all other values are interpreted as a request for a close ended TBF (i.e., the
RLC_Octet_Count value indicates the number of blocks the MS must transmit).
The RLC_Octet_Count field is also used to change the priority between uplink and
downlink, such that the uplink allocation is extended.
When the MS asks for uplink resources, if RLC_Octet_Count=0 or if RLC_Octet_Count
is more than a value defined by the THSULBAL parameter, then a switch from downlink
priority to uplink priority is executed; in this case, the number of blocks the MS must
transmit is supposed to be quite big.
Otherwise a timer defined by TSULBAL is activated. If the uplink TBF is closed until the
timer is running, then the timer is stopped and downlink priority in maintained; if the timer
elapses and the uplink TBF is still opened, a switch from downlink priority to the uplink
one is executed.

9.9.6 Suspend/Resume Procedures


These procedures are used when the Circuit Switched dedicated mode is entered and
it is temporarily needed to suspend GPRS/EGPRS service.
In fact, when a GPRS/EGPRS-attached MS enters the Circuit Switched dedicated
mode, and when the MS limitations make it unable to handle both the CS dedicated
mode and the packet switched transfer mode, the MS requests the network to suspend
PS services.
This is the case of an MS operating in Class B mode. The MS is attached to both the PS
and other CS services, but it can only operate one set of services at a time (see
"9.1 Mobile Stations for Packet Switched Services").

226 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

For instance, an ongoing downlink transmission can be suspended, a circuit switched


call accepted and, after the call release, the packet switched data transmission
continued.
The suspend procedure is explained in the following (see Fig. 9.16):
1. The MS enters the CS dedicated mode.
2. The MS sends a GPRS SUSPENSION REQUEST message to the BSC to inform
the BSC that it will suspend PS services. The GPRS SUSPENSION REQUEST
message contains the following:
– Temporary Logical Link Identity (TLLI)
– Routing Area Identifier (RAI)
– Suspension Cause (SC)
The BSC stores the information related to the request, namely the TLLI and the RAI,
associating them with the CS call of the MS.
3. When the BSC receives the GPRS SUSPENSION REQUEST message, it sends a
SUSPEND message to the SGSN containing the TLLI and the RAI; when the
message is sent, the T3 timer is started.
If the T3 timer expires without receiving the ACK message from the SGSN, the
SUSPEND message will be sent again and the T3 timer is restarted. This retry step
is repeated up to SUSPEND-RETRIES times (SUSPEND-RETRIES=3.)
If the T3 timer expires SUSPEND-RETRIES times or a SUSPEND NACK is received
from the SGSN, the suspend procedure is considered unsuccessful and an Alarm
Reporting notification is sent to LMT/RC.
4. If the SGSN acknowledges the SUSPEND message by returning the SUSPEND
ACK one, while the T3 timer is running, the procedure is considered successful. The
BSS will store TLLI and RAI in order to be able to request the SGSN to resume PS
services when the MS leaves dedicated mode; the BSC also receives the SRN
(Suspend Reference Number) information from the SGSN.
The T3 timer is stopped and the involved TLLI is marked as "Suspended" in the BSC.
Each received SUSPEND-ACK message is discarded (without notification towards
i RC/LMT) by the BSC either if the MS related to the received TLLI/RAI is already
"Suspended" or if the received TLLI/RAI does not correspond to an MS requiring the
suspension.

The BSC will suspend the GPRS/EGPRS service for the relevant MS, meaning that
i no traffic for the MS (TLLI/RAI) will be forwarded to the MS, even if the radio
resources are kept allocated to be available for the following Resume.

A30808-X3247-L24-5-7618 227
GPRS/EGPRS Global Description Information
System

MS BSC SGSN MSC/VLR

Dedicated Mode is entered

GPRS SUSPENSION REQUEST

SUSPEND (TLLI, RAI)


Start T3

Stop T3 SUSPEND ACK (TLLI, RAI, SRN)

Fig. 9.16 Suspend Procedure

The BSC will start the resume procedure as soon as the circuit switched dedicated mode
is left, that is when the MS is disconnected from the MSC.
Two cases must be distinguished:
a) during the suspension period, the MS has remained in the same cell;
b) during the suspension period, the MS has changed cell or routing area.
In case a), the resumption request is managed with the SGSN; in the following this
procedure is explained (the procedure in case of successful resume is shown in
Fig. 9.17):
1. The BSC starts the resume procedure sending the RESUME message containing
the TLLI, the RAI, and the SRN towards the SGSN; the T4 timer is also started.
If the T4 timer expires without receiving the ACK message from the SGSN, the
RESUME message will be sent again and the T4 timer is restarted. This retry step
is repeated up to RESUME-RETRIES times (RESUME-RETRIES=3).
In case the T4 timer expires RESUME-RETRIES times or a RESUME NACK is
received from the SGSN, the resume procedure is considered unsuccessful and an
Alarm Reporting notification is sent to the LMT/RC.
2. If the SGSN acknowledges the RESUME message by returning the RESUME ACK
one while the T4 timer is running, the procedure is considered successful and the
T4 timer is stopped.
3. In both cases, successful and unsuccessful, the involved TLLI is marked as "Not
Suspended" in the BSC. Moreover, the BSC removes the information related to the
previous suspend request and, in both cases, it closes the procedure by sending a
CHANNEL-RELEASE message to the MS with the following topics:
– in the successful case, the message includes the "GPRS Resumption" info-
element set to "resumption of PS services successfully acknowledged";

228 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

– otherwise, the "GPRS Resumption" will be set to "resumption of PS services not


successfully acknowledged"
In the former case, the MS will consider the GPRS/EGPRS service resumed, in the
latter, it will be invited to initiate a Routing Area Update procedure.
Each received RESUME-ACK message is discarded if either the MS related to the
i received TLLI/RAI is already "Resumed" or the received TLLI/RAI does not corre-
spond to a MS for which the resumption has been required.

4. Eventually, the BSS determines that the circuit-switched radio channel will be
released. If the BSS is able to request the SGSN to resume PS services, the BSS
will send a Resume (TLLI, RAI) message to the SGSN. The SGSN acknowledges
the successful outcome of the resume by returning Resume Ack;
5. The BSS sends an RR Channel Release (Resume) message to the MS. Resume
indicates whether the BSS has successfully requested the SGSN to resume
GPRS/EGPRS services for the MS, i.e., whether Resume Ack was received in the
BSS before the RR Channel Release message was transmitted. The MS leaves
dedicated mode;
6. If the BSS did not successfully request the SGSN to resume PS services, or if the
RR Channel Release message was not received before the MS left dedicated mode,
then the MS will resume GPRS/EGPRS services by sending a Routing Area Update
Request message to the SGSN, as described in subclause "Routing Area Update
Procedure".

MS BSC SGSN MSC/VLR

RESUME (TLLI, RAI, SRN)


Start T4

RESUME ACK (TLLI, RAI)


Stop T4

Channel Release

Fig. 9.17 Resume Procedure (the MS has remained in the same cell - Successful
Resume)

In case b), the resume procedure towards the SGSN is skipped; in the following, this
procedure is explained (the procedure is shown in Fig. 9.18):
1. The BSC removes the information related to the previous suspend request and it
immediately sends a CHANNEL-RELEASE message to the MS. The information
element "GPRS Resumption" will not be included in the message.

A30808-X3247-L24-5-7618 229
GPRS/EGPRS Global Description Information
System

2. As a result, if the Routing Area was changed, a Routing Area Update procedure is
initiated by the MS; if the cell was changed but not the Routing Area, depending on
its state, the MS will continue in the following way:
– Ready state: a Cell Update procedure is initiated by the MS. The SGSN is aware
of the cell to which the MS currently belongs.
– Standby state: the MS does nothing. When the SGSN side Ready Timer expires,
the SGSN will page the MS in the Routing Area it knows, to find the right cell.

MS BSC SGSN MSC/VLR

Channel Release
(no Resumption Result)

Routing Area Update Request

Fig. 9.18 Resume Procedure (The MS has changed the Routing Area)

If the MS performs an inter-BSC handover while suspended, the old BSC transfers the
“Old BSS to New BSS information” IE to the new one; this information element contains
the “GPRS Suspend information” field (this field contains the SUSPEND ACK PDU
message sent on the Gb interface.
With this information the new BSC is able to resume the MS that was suspended in the
old BSC, without executing any routing area update procedure.
In BR7.0, the GPRS/EGPRS Suspend/Resume feature is automatically enabled in the
i system, and both T3 and T4 timers cannot be set by the operator, but assume the
following default values:
- T3 = 5 s
- T4 = 1 s

9.9.7 Notes About GPRS/EGPRS TBF Scheduling


The distribution of the active TBFs over the available GPRS/EGPRS carriers, and the
multislot allocation of a particular TBF on the timeslots of the corresponding
GPRS/EGPRS carrier, is done by the resource allocation algorithm, which is described
in "5.3 Management of Packet Data Channels".
Once resources have been allocated, allowing each TBF to reach the maximum
required throughput, it is up to the scheduler to dynamically assign permissions to
access the physical channels, when several TBFs are multiplexed on the same
resources.
Therefore, the scheduler is able to handle the case when EGPRS and GPRS mobiles
are multiplexed on the same PDCHs: in this case, the main problem is due to the fact

230 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

that GPRS mobiles are not able to read the USF in downlink blocks transmitted with
8PSK modulation; therefore, uplink and downlink scheduling must be performed jointly,
trying to avoid setting USFs for GPRS mobiles in 8PSK coded downlink blocks.
Moreover, when different TBFs are multiplexed together, the scheduler takes into
account their different QoS requirements. The scheduler then assures that each TBF is
served according to its own priority.

9.9.7.1 Supported QoS Attributes


As previously mentioned, the task of the scheduler is to distribute permissions to access
the physical resources, after the allocation phase performed by the resource manager.
The QoS requirements to fulfil are linked to the R97/98 QoS attributes that can be
handled by the BSS; these attributes are different between the uplink and downlink
directions.
Regarding downlink scheduling, the BSC handles the following two R97/98 attributes
specified in the DL-UNITDATAs coming from the Gb interface:
– Peak Throughput (used in the allocation phase)
– Service Precedence
The Service Precedence, strictly speaking, indicates the relative importance of main-
taining the service commitments under abnormal conditions, for example which packets
are discarded in the event of problems such as limited resources or network congestion.
Therefore, even if this attribute should be used to handle congestion cases, it seems
reasonable to simplify things, handling it in the scheduling phase.
The Service Precedence is then used to assign different priorities to different connec-
tions; there are three possible values, from 1 (highest priority) to 3 (lowest priority).
Consequently, the scheduling priorities (needed to prioritize among TBFs that share the
same resources) are defined as follows:

QoS Attribute Scheduling Priority


Service Precedence = 1 1
Service Precedence = 2 2
Service Precedence = 3 3
When multiple TBFs are allocated on the same physical resources, they will be served
according to their own scheduling priorities, see next paragraphs for details.

Regarding the scheduling of uplink TBFs, the Radio Priority attribute is used. Such an
attribute is mapped one to one to uplink scheduling priority as follows:

QoS Attribute Scheduling Priority


Radio Priority = 1 1
Radio Priority = 2 2
Radio Priority = 3 3
Radio Priority = 4 4

9.9.7.2 Scheduling Process


As a general rule, the scheduling algorithm for each PDCH checks the active TBFs onto
that timeslot, both in uplink and downlink directions; then it verifies the scheduling prior-

A30808-X3247-L24-5-7618 231
GPRS/EGPRS Global Description Information
System

ities of each TBF, and associates the corresponding scheduling weights Wk to each
priority.
In fact, each scheduling priority is associated with a specific scheduling weight: the
association between priorities and weights can be performed by the user with the
following parameters:
– SCHWEIPRI1: weight associated to scheduling priority 1
– SCHWEIPRI2: weight associated to scheduling priority 2
– SCHWEIPRI3: weight associated to scheduling priority 3
– SCHWEIPRI4: weight associated to scheduling priority 4
For each direction of transmission and for each timeslot, the algorithm selects a TBF
using an approach that guarantees that each TBF(i) is selected W(i) times each
sum(Wk) extractions, where W(i) is the scheduling weight of TBF(i) and sum(Wk) is the
sum of the scheduling weights of all TBFs allocated on the same timeslot.
Besides, there are several cases that require high priority handling, both in downlink and
uplink directions. These cases are as follows:
• Polling Requests: when an MS must be polled, the scheduler manages the down-
link block (containing the RRBP value) for this MS, with a higher priority than other
blocks to be sent in downlink direction;
• Downlink Control Blocks for uplink TBFs: when a downlink control block (i.e.,
Packet Uplink Ack/Nack message) must be sent for an uplink TBF, the scheduler
manages the downlink block for this uplink TBF, with a higher priority than other
blocks to be sent in downlink direction;
• Constraint due to PCCCH Scheduling: if the PCCCH channel is allocated on a
certain timeslot, some radio blocks in a multiframe should be reserved for PCCCH.
Therefore, the scheduler manages, on this timeslot, both the downlink PCCCH
blocks (PBCCH, PAGCH, PPCH) and the uplink PCCCH blocks (PRACH) with a
higher priority than other blocks to be sent in downlink/uplink direction on the same
timeslot.
Besides, when EGPRS and GPRS mobiles are multiplexed on the same PDCHs some
additionally constraints must be taken into account. The main problem is due to the fact
that GPRS mobiles are not able to read the USF in downlink blocks transmitted with
8PSK modulation. Therefore, uplink and downlink scheduling must be performed jointly,
trying to avoid setting USFs for GPRS mobiles in 8PSK-coded downlink blocks.
In this case, the problem arises when, e.g., a downlink block coded with 8PSK modula-
tion must be sent and at the same time the USF should be coded with GMSK modulation
allowing to a GPRS-only mobile station to read the USF value and transmit on the next
uplink block. To solve this incompatibility, the following approach is used:
a) if the downlink block corresponds to a TBF that uses the GMSK modulation, an USF
(the first in the list of the USF to be transmitted) corresponding to a GMSK mobile is
selected; if there are no GMSK USF, then an 8PSK one is selected, because this
choice does not cause any problem;
b) If the downlink block corresponds to a TBF that uses the 8PSK modulation, an USF
(the first in the list of the USF to be transmitted) corresponding to a 8PSK mobile is
selected; if there are no more 8PSK USF to be scheduled (in the list), then:
– one more 8PSK USF is selected (starting from UL TBFs with higher
priority/weight)
– at the same time one “GMSK USF” is cancelled from the list
– after a few cancellations for a given TBF, a GMSK USF for that TBF is inserted in
an High_priority list (i.e., a list of TBF to be served with an higher priority). This

232 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

guarantees that, even in worst case scenarios (only 8PSK TBFs in downlink),
some GMSK USFs are still transmitted.

A30808-X3247-L24-5-7618 233
GPRS/EGPRS Global Description Information
System

10 GPRS/EGPRS Functionalities

10.1 Cell Selection and Re-selection


No Handover functionality is foreseen for PS services: the MS selects the best cell
following cell re-selection criteria.
If the MS is involved in data transfer, packets may be lost during cell re-selection. Upper
layers will then recognize the inconsistency, discard the frame and ask for a retransmis-
sion.
In GPRS standby and ready states (see "9.3.1 Mobility Management States"), cell re-
selection is performed by the MS.
The only exception regards class A mobile stations in dedicated mode of a circuit
switched connection: in this case the cell is determined by the network according to the
handover procedures, since the handover takes precedence over GPRS/EGPRS cell
re-selection. When the circuit switched connection is released, the MS resumes the cell
re-selection process.
For GPRS/EGPRS mobile stations, new re-selection criteria (C1, C31 and C32) can be
used. These new algorithms apply to the MSs attached to GPRS/EGPRS if the PBCCH
exists in the serving cell.
The C1 criterion is the same criterion used in GSM cell selection and re-selection
i processes, but it makes use of new parameters (see "10.1.2.1 GPRS/EGPRS Path
Loss Criterion (C1 Criterion)").

If PBCCH exists, the MS is not required to monitor system information on both the
serving cell and non-serving cells, but is only required to monitor system information on
PBCCH of the serving cell. In other words:
• if PBCCH is configured, the GPRS/EGPRS MS retrieves all of the information,
regarding both the serving cell and neighboring cells, from the serving PBCCH; the
MS monitors the other BCCH carriers only to take signal level measurements;
• if PBCCH is not configured, the GPRS/EGPRS MS retrieves all of the information
regarding the serving cell from the serving BCCH, while the information about neigh-
boring cells are taken from the BCCH carriers of the neighboring cells; the MS also
monitors the other BCCH carriers to take signal level measurements.
If PBCCH is not configured, i.e., PS services are supported only on BCCH, “old” C1 and
C2 criteria are used for cell selection and re-selection purposes.
In addition, it is possible to run a procedure, which is called Network Controlled Cell-
Reselection (see "10.3 Network Controlled Cell Reselection and Traffic Control
Management"), where the network may control the cell selection process.
If the PBCCH is configured, cells to be monitored for cell re-selection are defined in the
BA(GPRS) list, which is broadcast on PBCCH. This list could be different from the
BA(BCCH) list, that is used for GSM. If PBCCH does not exist, BA(GPRS) list is equal
to the BA(BCCH) one (see "10.1.4 Management of GPRS/EGPRS Neighboring Cells").

10.1.1 Measurements for Cell Selection and Re-selection


The MS measures the received RF signal level on the following:
– BCCH carrier of the serving cell
– BCCH carriers of surrounding cells as indicated in the BA(GPRS) list

234 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Then it calculates the average received level (RLA_P) for each carrier. In addition the
MS verifies the BSIC of the BCCH carriers. Only cells with allowed BSIC are considered
for re-selection purposes.
A distinction must be done between mobile stations in packet idle mode and mobile
stations in packet transfer mode:
a) Packet Idle Mode: whilst in packet idle mode a MS monitors continuously all BCCH
carriers as indicated by the BA(GPRS) list and the BCCH carrier of the serving cell.
At least one receive signal level measurement sample on each BCCH carrier is
taken for each paging block monitored by the MS according to its current DRX mode
and its paging group (see "9.8.3.2 Discontinuous Reception").
The MS will take at least one measurement for each BCCH carrier for every 4
seconds. The MS is not required to take more than 1 sample per second for each
BCCH carrier. RLA_P is an average level determined using samples collected over
a period of 5 s, and is maintained for each BCCH carrier. The same number of
measurement samples is taken for all BCCH carriers, and the samples allocated to
each carrier will as far as possible be uniformly distributed over the evaluation
period. At least 5 received signal level measurement samples are required for a valid
RLA_P value. The list of the 6 strongest non serving carriers are updated at a rate
of at least once per running average period.
The MS will attempt to check the BSIC for each of the 6 strongest non serving cell
BCCH carriers at least every 14 consecutive paging blocks of that MS, or 10
seconds, whichever is greater. If a change of BSIC is detected then the carrier will
be treated as a new carrier.
In the case of a multiband MS, the MS will attempt to decode the BSIC, if any BCCH
carrier with unknown BSIC is detected among the number of strongest BCCH
carriers in each band, as indicated by the GNMULBAC parameter
(MULTIBAND_REPORTING); this parameter is broadcast on PBCCH, or if PBCCH
does not exist, on BCCH.
b) Packet Transfer Mode: while in packet transfer mode, a MS continuously monitors
all BCCH carriers as indicated by the BA(GPRS) list and the BCCH carrier of the
serving cell. In every TDMA frame, a received signal level measurement sample is
taken on at least one of the BCCH carriers, one after the another. RLA_P is an
average value determined using samples collected over a period of 5 s, and is main-
tained for each BCCH carrier. The samples allocated to each carrier will as far as
possible be uniformly distributed over the evaluation period. At least 5 received
signal level measurement samples are required for a valid RLA_P value. The MS will
attempt to check the BSIC for each of the 6 strongest non serving cell BCCH carriers
as often as possible, and at least every 10 seconds.
A multi-RAT MS is allowed to extend this period to 13 seconds, if the neighbor cell
list contains cells from other RATs and if indicated by the GUMTSSRHPRI param-
eter.
The MS will use the two Idle frames of the PDCH multiframe for this purpose. These
frames are termed “search” frames.
In the case of a multiband MS, the MS will attempt to decode the BSIC, if any BCCH
carrier with unknown BSIC is detected, among the number of strongest BCCH
carriers in each band, as indicated by the GNMULBAC parameter
(MULTIBAND_REPORTING); this parameter is broadcast on PBCCH, or if PBCCH
does not exist, on BCCH.

A30808-X3247-L24-5-7618 235
GPRS/EGPRS Global Description Information
System

10.1.2 Cell selection and Re-selection Criteria

10.1.2.1 GPRS/EGPRS Path Loss Criterion (C1 Criterion)


The MS measures the received signal level on the PBCCH carriers of the serving cell
and the surrounding cells and calculates the mean received level (RLA_P) for each
carrier, where:
– RLA_P(s) is the averaged level for the serving cell
– RLA_P(n) are the averaged levels for neighboring cells
Cells to be monitored for cell reselection are defined by the BA(GPRS) list, which is
broadcast on PBCCH. At least 5 received signal level measurement samples are
required for a valid RLA_P:

RLA_P = 1/5 * (GPRS_RXLEV1 + GPRS_RXLEV2 + ...+ GPRS_RXLEV5)

The path loss criterion C1, i.e., the minimum signal level criterion for GPRS/EGPRS cell
selection and cell re-selection, is defined by the following:

C1 = RLA_P – GPRS_RXLEV_ACCESS_MIN – Max


(0, GPRS_MS_TXPWR_MAX_CCH–P)

Where:

- P is the Power Class of the MS


- GPRS_RXLEV_ACCESS_MIN is the minimum allowed received level to access a
cell; the user can define this value with the GRXLAMI parameter
- GPRS_MS_TXPWR_MAX_CCH is the maximum power that the Mobile Station can
use to access the cell; the user can define this value when configuring the
GMSTXPMAC parameter related to the Managed Object PTPPKF. This attribute spec-
ifies the maximum TX power level that a Mobile Station may use when accessing the
system in presence of the PBCCH. In case the PBCCH does not exist the Mobile
Station uses the MSTXPMAXCH attribute to evaluate the path loss criterion parameter
“C1” whereas the BSC uses the GMSTXPMAC attribute for the evaluation of the same
parameter. For the serving and neighbour cells the GMSTXPMAC value is sent in
broadcast on the PBCCH of the serving cell.

The path loss criterion is satisfied if C1>0.


This means that the minimum allowed received downlink level to access the network
must be above a threshold defined by GRXLAMI value.
To ensure a sufficient uplink received level even for MS of low transmit power level P,
the C1 criteria works as follows:

If P < GPRS_MS_TXPWR_MAX_CCH
C1 = RLA_P – GPRS_RXLEV_ACCESS_MIN – (GPRS_MS_TXPWR_MAX_CCH– P)

236 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

i.e., the received level must be higher than the access threshold
(GPRS_RXLEV_ACCESS_MIN) plus another term, given by the difference between
the maximum power that can be transmitted in the cell
(GPRS_MS_TXPWR_MAX_CCH) and the nominal power of the MS (P).

If P > GPRS_MS_TXPWR_MAX_CCH
C1 = RLA_P – GPRS_RXLEV_ACCESS_MIN

i.e., the received level must only be higher than the access threshold
(GPRS_RXLEV_ACCESS_MIN); in this case, the nominal power of the MS is higher
than GPRS_MS_TXPWR_MAX_CCH.

C1 criterion is an assessment about the field strengths (on both uplink and downlink
i directions).
If PBCCH is used, the C1 criterion is calculated by the same formula used in GSM, but
with a separate parameter set (i.e., GRXLAMI and GMSTXPMAC), which is transmitted
on the PBCCH. With this separate parameter set, it is possible for the network operator
to configure, in a different way, the cell selection and reselection procedures, for
GPRS/EGPRS and not-GPRS/EGPRS subscribers.
If PBCCH is not configured, the C1 criterion is calculated by means of the same formula
and parameters (i.e., RXLEVAMI and MSTXPMAXCH) used for GSM cell selection and
re-selection.
Please remember that on PBCCH the network has the chance to indicate in the
BA(GPRS) list a different set of neighboring cells with respect to the BA list transmitted
on BCCH (see 10.1.4).

Beside the C1 radio criterion, there are some other criteria for a cell to be suitable for
GPRS/EGPRS cell selection purpose: a cell is considered suitable for GPRS/EGPRS
cell selection if:
1. C1 is greater than 0
2. the cell belongs the selected PLMN
3. the cell supports PS services
4. the cell is not barred

10.1.2.2 C31 Criterion


The C31 signal level threshold criterion for hierarchical cell structures (HSCs) is used to
determine whether prioritized hierarchical GPRS/EGPRS cell re-selection is applied. It
is defined by:

Serving Cell
C31(s) = RLA_P(s) – HCS_THR(s)

Neighboring Cell
PRIORITY_CLASS(n) = PRIORITY_CLASS (s)
C31(n) = RLA_P(n) – HCS_THR(n)

A30808-X3247-L24-5-7618 237
GPRS/EGPRS Global Description Information
System

Neighboring Cell
PRIORITY_CLASS(n) < > PRIORITY_CLASS (s)
There are two cases:
If T<=GPRS_PENALTY_TIME
C31(n) = RLA_P(n) – HCS_THR(n) – GPRS_TEMPORARY_OFFSET(n)

If T > GPRS_PENALTY_TIME
C31(n) = RLA_P(n) – HCS_THR(n)

Where:
• HCS_THR is the signal threshold for applying GPRS/EGPRS hierarchical cell struc-
tures criteria in cell reselection. The user can define this threshold using the
GHCSTH parameter. The user defines the threshold both for the cell and for its
neighboring cells, in fact:
– HCS_THR(s) represents the threshold of the serving cell; the user specifies it by
the GHCSTH parameter of the PTPPKF object;
– HCS_THR(n) represents the thresholds of neighboring cells; the user sets a
HCS_THR(n) value for every adjacent relationship, by the GHCSTH parameter of
the ADJC object;
• PRIORITY_CLASS is the priority of each cell. The user can define this priority by the
GHCSPC parameter (a higher value means a higher priority). The user defines the
priority both for the cell and for its neighboring cells, in fact:
– PRIORITY_CLASS(s) represents the priority of the serving cell; the user specifies
it by the GHCSPC parameter of the PTPPKF object;
– PRIORITY_CLASS(n) represents the priority of neighboring cells; the user sets a
PRIORITY_CLASS(n) value for every adjacent relationship, with the GHCSPC
parameter of the ADJC object;
• GPRS_TEMPORARY_OFFSET(n) applies a negative offset to C31 for the duration
of GPRS_PENALTY_TIME(n) after the timer T has started for that cell.
The T timer is started in the MS for each cell in the list of the 6 strongest neighboring
cells, as soon as it is placed in the list. T is reset to 0 if the cell is removed from the
list.
GPRS_PENALTY_TIME is the duration for which GPRS_TEMPORARY_OFFSET
applies.
The user sets a GPRS_TEMPORARY_OFFSET(n) value and a
GPRS_PENALTY_TIME(n) value for every adjacent relationship, by GTEMPOFF
and GPENTIME parameters of the ADJC object.
Regarding the previous parameters, it is important to underline that their values are
broadcasted on the PBCCH of the serving cell, i.e., the MS can retrieve all of the cell re-
selection information from the PBCCH of the serving cell without monitoring the other
neighboring carriers. To understand how this feature is implemented, refer to
"10.1.4 Management of GPRS/EGPRS Neighboring Cells". This is different from the
traditional GSM implementation for which the MS must retrieve the cell re-selection
parameters of the neighboring cells, by reading their BCCH carriers.

238 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

C31 is used for hierarchical cell structures; the advantage is that C31 also uses a priority
i mechanism. It is necessary to introduce C31 into GPRS/EGPRS, to make re-selection
for PS services similar to the GSM handover algorithm.
The MS needs to get information of the neighbor cells (e.g., in which layer the neigh-
boring cells are laying, and the priority of the neighbor cells), to decide about cell re-
selection. For CS services, the Handover decision is done completely by the BTS, so it
is not necessary to give additional information to the MS.

10.1.2.3 C32 Criterion


The C32 cell ranking criterion is used to select cells among those with the same priority.
It is defined by:

Serving Cell
C32(s) = C1(s)

Neighboring Cell
PRIORITY_CLASS(n) = PRIORITY_CLASS (s)
There are two cases:
If T <= GPRS_PENALTY_TIME
C32(n) = C1(n) + GPRS_RESELECT_OFFSET(n) –
GPRS_TEMPORARY_OFFSET(n)

If T > GPRS_PENALTY_TIME
C32(n) = C1(n) + GPRS_RESELECT_OFFSET(n)

Neighboring Cell
PRIORITY_CLASS(n) < > PRIORITY_CLASS (s)
C32(n) = C1(n) + GPRS_RESELECT_OFFSET(n)

Where:
• PRIORITY_CLASS is the priority of each cell. The user can define this priority by the
GHCSPC parameter (a higher value means a higher priority). The user defines the
priority both for the cell and for its neighboring cells, in fact:
– PRIORITY_CLASS(s) represents the priority of the serving cell; the user specifies
it by the GHCSPC parameter of the PTPPKF object;
– PRIORITY_CLASS(n) represents the priority of the neighboring cells; the user
sets a PRIORITY_CLASS(n) value for every adjacent relationship, by the
GHCSPC parameter of the ADJC object;
• GPRS_RESELECT_OFFSET(n) is a positive offset that increases the priority of cell
in the list of the strongest neighbor cells. The user sets a
GPRS_RESELECT_OFFSET(n) value for every adjacent relationship, by
GRESOFF parameter of the ADJC object;

A30808-X3247-L24-5-7618 239
GPRS/EGPRS Global Description Information
System

• GPRS_TEMPORARY_OFFSET(n) applies a negative offset to C32 for the duration


of GPRS_PENALTY_TIME(n) after the timer T has started for that cell.
The T timer is started in the MS for each cell in the list of the 6 strongest neighboring
cells, as soon as it is placed on the list. T is reset to 0 if the cell is removed from the
list.
GPRS_PENALTY_TIME is the duration for which GPRS_TEMPORARY_OFFSET
applies.
The user sets a GPRS_TEMPORARY_OFFSET(n) value and a
GPRS_PENALTY_TIME(n) value for every adjacent relationship, by GTEMPOFF
and GPENTIME parameters of the ADJC object.
GPRS_RESELECT_OFFSET, GPRS_TEMPORARY_OFFSET, PRIORITY_CLASS
and GPRS_PENALTY_TIME are broadcast on PBCCH of the serving cell.
Regarding the previous parameters it is important to underline that their values are
broadcasted on the PBCCH of the serving cell, i.e., the MS can retrieve all of the cell re-
selection information from the PBCCH of the serving cell without monitoring the other
neighboring carriers. To understand how this feature is implemented, see
"10.1.4 Management of GPRS/EGPRS Neighboring Cells". This is different from the
traditional GSM implementation for which the MS must retrieve the cell re-selection
parameters of the neighboring cells, by reading their BCCH carriers.
C32 is similar to the C2 criteria used for GSM, but it makes use of GPRS/EGPRS
i parameters.
C32 contains, in addition to C1, an offset (to make a cell better or worse than another)
and a temporary offset, which is used to make the cell worse during the first x seconds
(i.e., the MS must "see" that cell for that period of time before it may re-select it; this can
be used to prevent some cells from beeing re-selected by a fast driving MS).

10.1.3 Cell Re-selection Algorithm


The MS makes a cell reselection if:
• C1 (serving cell) < 0 for a period of 5 seconds
• MS detects DL Signalling failure (e.g., no paging has been possible)
• Cell becomes barred
Beside these conditions that regard only the serving cell, other conditions derive from
the comparison between the serving cell and the neighboring ones.
There are two different conditions:
a) Both GPRS/EGPRS serving cell and GPRS/EGPRS neighboring cells config-
ured with BCCH;
b) Both GPRS/EGPRS serving cell and GPRS/EGPRS neighboring cells config-
ured with PBCCH.

a) GPRS/EGPRS serving cell and GPRS/EGPRS neighboring cells configured with


BCCH
C2 (GSM) criteria is switched on; a cell reselection is executed if:
C2 (GPRS/EGPRS serving cell) < C2 (suitable GPRS/EGPRS neighboring cell)
If the suitable neighboring cell is in the same location area for a period of 5sec.

240 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

C2 (GPRS/EGPRS serving cell) + CELL_RESELECT_HYST < C2 (suitable


GPRS/EGPRS neighboring cell)
If the suitable neighboring cell is in another location area for a period of 5 s.
The C1 value of the neighboring cell must obviously be greater than 0.
i

C2 (GSM) criteria is not switched on; a cell reselection is executed if:


C1 (GPRS/EGPRS serving cell) < C1 (suitable GPRS/EGPRS neighboring cell)
If the suitable neighboring cell is in the same location area for a period of 5 seconds.

C1 (GPRS/EGPRS serving Cell) + CELL_RESELECT_HYST < C1 (suitable


GPRS/EGPRS neighboring cell)
If the suitable neighboring cell is in another location area for a period of 5 s.
The CELL_RESELECT_HYST value is defined by the user through the GSM CELL-
i RESH parameter.

b) GPRS/EGPRS serving cell and GPRS/EGPRS neighboring cells configured with


PBCCH
First, the C31 criterion for the serving cell and all neighboring cells is calculated.
The best cells are found under all of these cells, therefore the following is checked:
If there are cells for which C31>=0, then the PRIORITY_CLASS is checked only for
these cells.
If there is only one cell with the highest PRIORITY_CLASS, then this cell will be the best
cell to make cell reselection on.
If there are more cells with the highest PRIORITY_CLASS, then the C32 criterion will be
calculated for these cells. The cell with the highest C32 criterion will be the best cell on
which to make cell reselection.
If there are not cells for which C31>0, then the C32 criterion is calculated for all cells.
The cell with the highest C32 criterion will be the best cell on which to make cell rese-
lection.

When evaluating the better cell, the following hysteresis values will be subtracted from
the C32 value of the neighboring cells:
a) in standby state, if the new cell is in the same routing area no hysteresis values are
subtracted;
b) in ready state, if the new cell is in the same routing area, a
GPRS_CELL_RESELECT_HYSTERESIS value is subtracted to delay a cell re-
selection, since a TBF might be interrupted; the user sets this hysteresis with the
GCELLRESH parameter.
If the C31H parameter (C31_HYST) is set to TRUE, and the MS is in MM
i ready state, the GPRS_CELL_RESELECT_HYSTERESIS is also
subtracted from the C31 value for the neighboring cells.

c) in standby or ready state, if the new cell is in a different routing area a


RA_RESELECT_HYSTERESIS value is subtracted to delay a cell re-selection,

A30808-X3247-L24-5-7618 241
GPRS/EGPRS Global Description Information
System

because routing area changes will produce a lot of extra signalling; the user sets this
hysteresis using the RARESH parameter.
d) if a cell re-selection occurred within the previous 15 seconds, a value of 5dB is
subtracted.

C31_HYST, RA_RESELECT_HYSTERESIS, and RA_RESELECT_HYSTERESIS are


broadcast on the PBCCH of the serving cell.
If the parameter C32QUAL (C32_QUAL) is set to TRUE, positive
i GPRS_RESELECT_OFFSET values are only applied to the neighbors with the highest
RLA_P value, among those cells for which C32 is compared above.

Cell re-selection for any other reason (see GSM 03.22) takes place immediately, but the
cell that the MS was camped on will not be returned to within 5 seconds, if another suit-
able cell can be found. If valid RLA_P values are not available, the MS will wait until
these values are available and then perform the cell re-selection if it is required. The MS
may accelerate the measurement procedure within the requirements to minimize the cell
reselection delay. If no suitable cells are found within 10 seconds, the cell selection algo-
rithm will be performed. Since information concerning a number of channels is already
known by the MS, it may assign high priority to measurements on the strongest carriers
from which it has not previously made attempts to obtain BCCH information and omit
repeated measurements on the known ones.

10.1.4 Management of GPRS/EGPRS Neighboring Cells

10.1.4.1 Handling of Neighboring Cells


A mechanism has been introduced to manage both internal adjacent cells (i.e., cells
belonging to the same BSC) and external adjacent cells (i.e., cells belonging to other
BSCs).
The management of both internal and external adjacent cells is provided by using the
TGTCELL attribute of the ADJC object; this attribute is a mandatory one that specifies
the path of the target cell instance (see the explanation below).
Upon creating an adjacent cell relationship, a distinction will be made between two
possible situations:
• adjacency between cells supporting only GSM service (adjacent relationships
between BTSs)
• adjacency between cells supporting GPRS/EGPRS service too (adjacent relation-
ships between PTPPKFs)

ADJACENT RELATIONSHIPS BETWEEN BTSs


In case of an adjacency to an internal BTS, the TGTCELL attribute will contain a refer-
ence (i.e., the complete path) to the internal target BTS.
In case of external adjacency, the TGTCELL attribute will contain a reference to a new
object, namely the TGTBTS object. A TGTBTS managed object instance contains a
copy of the attributes of an external BTS MOI, to which an adjacent relationship must be
made up. Once a TGTBTS MOI is configured, it will be treated by the system for the
management of the adjacencies, as the other internal target BTSs.

242 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

This means that, in the case in which the external target BTS is adjacent to more than
one internal serving BTS, it will no longer be necessary to replicate all of the attribute
values in every ADJC managed object instance, but it will be enough that the different
ADJC MOIs refer to the same TGTBTS MOI.
Fig. 10.1 shows the management of adjacent cells.

Fig. 10.1 Management of Adjacent Cells


Referring to BTS:1, two adjacent relationships are built:
– BTS:1/ADJC:1; internal adjacent relationship towards BTS:2 (belonging to BSC:1)
– BTS:1/ADJC:2; external adjacent relationship towards BTS:5 (belonging to BSC:2)
When the user creates the ADJC1 instance he/she must specify only those attributes
that are not enclosed in BTS:2 (i.e., handover management attributes), while for the
other attributes (i.e., BCCH, BSIC, CELLGLID, etc.) the TGTCELL attribute provides the
reference to BTS:2.
When the user creates the ADJC2 instance, he/she must specify only those attributes
that are not enclosed in BTS:5 (i.e., handover management attributes). The TGTCELL
attribute will provide the reference to the TGTBTS:0 instance, that contains all of the
attributes (i.e., BCCH, BSIC, CELLGLID, etc.) enclosed in BTS:5.
Therefore, the user must create the TGTBTS instance containing the attributes
belonging to the external cell, before creating an external adjacent relationship.

A30808-X3247-L24-5-7618 243
GPRS/EGPRS Global Description Information
System

ADJACENT RELATIONSHIPS BETWEEN PTPPKFs


The same considerations apply for the management of the adjacencies between
PTPPKFs. The TGTPTPPKF object class is introduced to configure, in the BSC data-
base, external GPRS/EGPRS neighboring cells.

Functional object Meaning

TGTPTPPKF Configure external GPRS/EGPRS neighboring


cells.

Tab. 10.1 TGTPTPPKF Object

The TGTPTPPKF object class is hierarchically dependent on the TGTBTS object class
in the containment tree. A TGTPTPPKF managed object instance will contain a copy of
the attributes, of an external PTPPKF instance, involved in the management of the adja-
cency. Once a TGTPTPPKF object instance is configured, it is treated by the system for
the management of the adjacencies, as the other internal target PTPPKFs.
Therefore:
• in case of internal adjacency, the TGTCELL attribute identifies the target PTPPKF
instance, through the reference to the superordinate BTS instance
• in case of external adjacency, the TGTCELL will identify the TGTPTPPKF object
instance, through the reference to the superordinate TGTBTS object instance
For each serving cell, it is possible to configure up to 32 neighboring cells supporting
i GPRS/EGPRS.

10.1.4.2 GPRS/EGPRS Neighboring Cells and Involved Parameters


As described in the previous chapters, cell re-selection parameters of both the serving
cell and its neighboring cells are transmitted on the PBCCH of the serving cell. In this
way, a MS camped on a cell can read all of the re-selection parameters without synchro-
nizing to the other cells.
This happens only if the PBCCH is configured on the serving cell; if the PBCCH is not
configured on the serving cell:
1. “old” C1 and C2 GSM criteria, and “old” GSM parameters are used for cell re-selec-
tion;
2. the MS takes the re-selection parameters of the neighboring cells from the BCCHs
of those cells; i.e., the MS must synchronize to these cells to read their data.
To allow the MS to read from the PBCCH of the serving cell, the re-selection parameters
of neighboring cells, the involved parameters are specified either in the PTPPKF or in
the ADJC/TGTPTPPKF objects. Tab. 10.2 illustrates these parameters.

Siemens ETSI

GRXLAMI GPRS_RXLEV_ACCESS_MIN
GMSTXPMAC GPRS_MS_TXPWR_MAX_CCH
GHCSTH HCS_THR

Tab. 10.2 Parameters involved in the management of GPRS/EGPRS neighboring


cells

244 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Siemens ETSI

GHCSPC PRIORITY_CLASS
GRESOFF GPRS_RESELECT_OFFSET
GTEMPOFF GPRS_TEMPORARY_OFFSET
GPENTIME GPRS_PENALTY_TIME

Tab. 10.2 Parameters involved in the management of GPRS/EGPRS neighboring


cells

Among parameters shown in Tab. 10.2, a distinction must be made with the following:
a) GRXLAMI and GMSTXPMAC parameters are only cell parameters; i.e., they must
be defined only on a cell basis, using the PTPPKF object. Nevertheless, to allow the
transmission of the neighboring cell parameters in the packet system information of
the serving cell, they will also be defined for every adjacent relationship; they must
be defined in the TGTPTPPKF object, only if the adjacent cell does not belong to the
same BSC of the serving one;
b) GHCSTH and GHCSPC parameters are defined both on a cell basis and on the
adjacent relationship basis;
– they are defined on a cell basis in the PTPPKF object to define the cell values;
– for each neighboring cell of the involved cell, the two parameters are specified in
the ADJC object, to specifies the values of the neighboring cells.
c) GRESOFF, GTEMPOFF, and GPENTIME parameters regard only adjacent rela-
tionships; i.e., for each neighboring cell of the involved cell, the parameter values are
specified in the ADJC object to indicate the values of the neighboring cells.
To manage the previous features, the following parameters (belonging to the ADJC
object) are involved:
1. The GSUP parameter is meaningful only if the PBCCH is configured on the serving
cell. It enables the transmission of parameters of the neighboring cell to which it
refers, in the packet system information of the serving cell. Besides the GSUP
parameter can be set also in a different band than the BCCH one. leaving to the
operator choice whether enable GPRs on extended band or not. A The following
considerations apply to the GSUP parameter:
– if the PBCCH is not enabled in the serving cell, the GSUP in meaningless, but PS
sevices are active in the neighboring cell. So, if the GSUP attribute of one of its
neighboring cell is set to FALSE in the serving cell, the MS can re-select the cell
(with GSM C1 and C2) without any problem.
– if the PBCCH is enabled in the serving cell, the neighborhoods that will be consid-
ered in the BA(GPRS) list will be those cells for which the GSUP attribute has
been set to TRUE in the adjacent relationship.
2. TGTCELL when the cell is internal, this parameter allows making a link to the BTS
(PTPPKF) object that defines this cell in the database; if the cell is external, this
parameter allows to making a link to the TGTBTS (TGTPTPPKF) object that defines
this cell in the database.

10.1.4.3 Configuration of an Adjacent Cell with GSUP= TRUE


When the PBCCH is configured on the serving cell and the user configures a neigh-
boring cell with GSUP =TRUE two cases exists:

A30808-X3247-L24-5-7618 245
GPRS/EGPRS Global Description Information
System

1. the neighboring cell is internal


2. the neighboring cell is external
In the following, these two possibilities are examined from the point of view of the
GPRS/EGPRS parameters shown in Tab. 10.2.

The Neighboring Cell is INTERNAL


If the target cell is internal, the user, with the TGTCELL attribute, executes a link to the
BTS object (and as a consequence to the PTPPKF one) related to this cell.
Regarding parameters shown in Tab. 10.2, the following considerations can be made:
1. GRXLAMI and GMSTXPMAC parameters must not be defined in the TGTPTPPKF
object (because this object, in this case, does not exist for the neighboring cell),
since they are cell parameters and are directly taken from the linked PTPPKF object;
2. GHCSPC and GHCSTH parameters could be specified in the ADJC object; if they
are not specified, their default values are taken;
3. GPENTIME, GRESOFF, and GTEMPOFF parameters could be specified in the
ADJC object; if they are not specified, their default values are taken.

The Neighboring Cell is EXTERNAL


If the target cell is external, the user, with the TGTCELL attribute, executes a link to the
TGTBTS object (and as a consequence to the TGTPTPPKF one) related to this cell,
since this object does not belong to the same BSC.
Regarding the parameter shown in Tab. 10.2, the following considerations can be
made:
1. GRXLAMI and GMSTXPMAC parameters must be defined in the TGTPTPPKF
object; they must have the same values that have in the PTPPKF object of the BSC
database where they have been defined;
2. GHCSPC and GHCSTH parameters could be specified in the ADJC object; if they
are not specified, their default values are taken;
3. GPENTIME, GRESOFF, and GTEMPOFF parameters should be specified in the
ADJC object; if they are not specified, their default values are taken.
When the target cell is external, both RACODE and RACOL parameters (see
i "9.2 Network Structure") must be specified in the TGTPTPPKF object; they must have
the same values that they have in the database where they have been defined.

10.1.4.4 Configuration of an Adjacent Cell with GSUP= FALSE


Since the GSUP attribute is set to FALSE for the neighboring cell, the GPRS/EGPRS
re-selection parameters of the neighboring cell are not transmitted in the serving cell.
Besides, this cell is not included in the BA(GPRS) list, i.e., the list of cells over which
GPRS/EGPRS re-selection can be done; from the GSM point-of-view, the cell is always
considered for cell re-selection purposes (using GSM C1 and C2 criteria).
To understand, consider the following example referred to a cell (in which the PBCCH
is configured) that has 10 neighboring cells. If:
– for 6 neighboring cells, the GSUP attribute is set to TRUE;
– for the remaining 4 cells, the GSUP attribute is set to FALSE;
then starting from the serving cell:
– 6 cells can be re-selected from the PS/CS services point-of-view by means of C1
(with GPRS/EGPRS parameters), C31 and C32 criteria;

246 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

– 4 cells can be re-selected only from GSM mobile stations by means of C1 (with GSM
parameters) and C2 criteria.
In this case, we have a BA(GPRS) list containing 6 cells, and a BA(GSM) list containing
10 cells.
Considering the previous example, if the PBCCH is not configured in the same cell, both
i the BA(GPRS) list and the BA(GSM) list are comprised of 10 neighboring cells, since
the GSUP attribute is meaningless.

When the PBCCH is configured on the serving cell and the user configures a neigh-
boring cell with GSUP = FALSE, independently if the neighboring cell is internal or
external, the GPRS/EGPRS re-selection parameters must not be specified in the ADJC
object, since they are not transmitted on the serving cell.

10.1.5 Abnormal Cell Re-selection


In the event of an “abnormal release with cell reselection” when PBCCH exists, an
abnormal cell reselection based on BA(GPRS) list is attempted.
For example the mobile station starts the “abnormal release with cell reselection” proce-
i dure after having made M+1 attempts to send a Packet Channel Request on PRACH,
without receiving any answer from the network (see "9.8.2.6 Uplink Access on PRACH
(Access Persistence Control)").

To enable the abnormal cell re-selection, the RAARET (RANDOM_ACCESS_RETRY)


parameter must be set to TRUE. This parameter allows enabling the abnormal cell rese-
lection starting from the serving cell, when an abnormal release with cell reselection
occurs.
The MS performs the following algorithm to determine which cell is to be used for this
cell reselection attempt:
1. the received level measurement samples, taken on the neighboring carriers indi-
cated in the BA (GPRS) list and received in the last 5 seconds, are averaged; the
carrier with the highest received average level (RLA) and with a permitted BSIC is
taken;
2. on this carrier, the MS attempts to decode the PBCCH data block containing the
parameters affecting cell selection;
3. if:
– the C1 parameter is greater than zero
– the cell belongs to the selected PLMN
– the cell is not barred
– access in another cell is allowed, i.e., RAARET parameter is set to TRUE,
the abnormal cell reselection is attempted on this cell;
4. if the MS is unable to decode the PBCCH data block, or if the conditions in step 3.
are not met, the carrier with the next highest received average level (RLA) and with
a permitted BSIC is taken; then the MS repeats steps 2. and 3.;
5. if the cells with the 6 strongest received average level values (and with permitted
BSICs) have been tried, but cannot be used, the abnormal cell reselection attempt
is abandoned, and the usual reselection algorithm (see "10.1.3 Cell Re-selection
Algorithm") is performed.
When an MS has executed an abnormal cell reselection, it is not allowed to reselect the
original cell for a number of seconds specified by the TRESEL parameter.

A30808-X3247-L24-5-7618 247
GPRS/EGPRS Global Description Information
System

The MS is under no circumstances allowed to access a cell to attempt abnormal cell


reselection later than 20 seconds after the detection within the MS of the abnormal
release causing the abnormal cell reselection attempt. In the case where the 20 seconds
elapses without a successful abnormal cell reselection, the attempt is abandoned and
the usual reselection algorithm (see "10.1.3 Cell Re-selection Algorithm") is performed.
In the event of an abnormal release with cell reselection when only BCCH exists, the MS
performs only the usual reselection algorithm, using the C1 and C2 criteria of GSM.

10.2 Cell Re-selection from GSM/GPRS/EGPRS Network to


UMTS Network
With the introduction of the UMTS network, it becomes very important to allow a dual
mode mobile station to re-select a UMTS cell starting from a GSM one.
Without this feature, once a dual mode GSM/UMTS terminal is camped on the
GSM/GPRS/EGPRS network, it does not have the possibility to select the UMTS
network.
Only the re-selection of a UMTS cell starting from the GSM network is considered; the
i opposite case is outside the scope of this chapter.

To allow this feature, the UMTS adjacent cell information must be sent (in the 3G Cell
Reselection List) on the broadcast carrier of the GSM network, to inform the UE/MS
which UMTS frequencies must be monitored for re-selection purposes.
For this monitoring, the MS may use search frames that are not required for BSIC
decoding. If indicated by the parameter GUMTSSRHPRI, the MS may use up to 25
search frames per 13 seconds without considering the need for BSIC decoding in these
frames.
According to both the type of service that the MS supports and the configuration of the
serving GSM cell, two different algorithms are defined to reselect a UMTS cell (either
FDD cell or TDD one), starting from a GPRS/EGPRS one; so we can have:
1. re-selection of the UMTS cell in case of circuit switched modality; this type of re-
selection is executed when:
– the MS is not GPRS/EGPRS attached (so it must use the circuit switched modality
to re-select UMTS cells);
– the MS is attached to GPRS/EGPRS services, but the PBCCH channel is not
configured in the GSM serving cell;
2. re-selection of the UMTS cell in case of packed switched modality; this modality is
used when the MS is GPRS/EGPRS attached and the PBCCH has been configured
in the serving cell.
The two cases are briefly discussed, in the following sections.

10.2.1 GSM-UMTS Re-selection Algorithm: Circuit Switched Case


When the PBCCH is not configured in the serving cell, the MS performs a cell re-selec-
tion to an adjacent UMTS (FDD or TDD) cell, only if the following conditions are satisfied
for a period of 5 seconds:

for the serving cell:


RSCP(UTRAN cell) >= RLA_C_s + XXX_Qoffset

248 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

for all of the suitable GSM neighboring cells:


RSCP(UTRAN cell) >= RLA_C_n + XXX_Qoffset

and also (only for FDD cells):


Ec/No (UTRAN FDD cell) >= FDD_Qmin

where:
– RSCP (Received Signal Code Power): is the power level received from the UMTS
cell
– Ec/No: is the signal/noise ratio regarding the UMTS FDD cell
– RLA_C_s: is the power level received from the serving cell
– RLA_C_n: is the power level received from neighboring cells
– XXX_Qoffset: offset for cell reselection for UMTS cells; the user sets this value by
the FDDQO parameter (BTS object) for FDD cells, or by TDDQO parameter (BTS
object) for TDD cells
– FDD_Qmin: minimum threshold for Ec/No for UMTS FDD cell re-selection; the user
sets this value with the FDDQMI parameter of the BTS object
If the 3G Cell Reselection list (sent by the network to the MS) includes UTRAN frequen-
cies, the MS will, at least every 5 seconds update the RLA_C value for the serving cell
and each of the at least 6 strongest non-serving GSM cells.
The MS will then reselect a suitable UTRAN cell if its measured RSCP value exceeds
the value of RLA_C for the serving cell and all of the suitable non-serving GSM cells by
the value XXX_Qoffset for a period of 5 seconds, and (only in case of FDD cells) the
UTRAN cells measured Ec/No value is equal or greater than the value FDD_Qmin. In
case of a cell reselection occurring within the previous 15 seconds, XXX_Qoffset is
increased by 5 dB.
If more than one UTRAN cell fulfills the above criteria, the MS selects the cell with the
highest RSCP value.
If the MS has reselected a GSM cell from an UTRAN one, cell reselection to UTRAN
does not occur within 5 seconds, if a suitable GSM cell can be found.
There is also a threshold by which the network indicates whether or not the measure-
ments for the cell reselection of the UMTS cells should be performed; the threshold indi-
cates if the signal level of the serving cell should be below or above it, in order to perform
UMTS cells measurements; the user sets this value by the QSRHI parameter of the BTS
object.
FDDQO, TDDQO, FDDQMI, and QSRHI parameters are broadcast on the BCCH of the
i serving cell.

10.2.2 GSM-UMTS Re-selection Algorithm: Packet Switched Case


When the PBCCH is configured in the serving cell, the MS performs a cell re-selection
to an adjacent UMTS (FDD or TDD) cell, only if the following three conditions are satis-
fied for a period of 5 seconds:

for the serving cell:


RSCP(UTRAN cell) >= RLA_P_s + XXX_GPRS_Qoffset

for all of the suitable GSM neighboring cells:

A30808-X3247-L24-5-7618 249
GPRS/EGPRS Global Description Information
System

RSCP(UTRAN cell) >= RLA_P_n + XXX_GPRS_Qoffset

and also (only for FDD cells):


Ec/No (UTRAN FDD cell) >= GFDD_Qmin

where:
– RSCP (Received Signal Code Power): it is the power level received from the UMTS
cell
– Ec/No: is the signal/noise ratio
– RLA_P_s: is the power level received from the serving cell
– RLA_P_n: is the power level received from the neighboring cells
– XXX_GPRS_Qoffset: offset for cell reselection for FDD cells; the user sets this value
with the FDDGQO parameter of the PTPPKF object for FDD cells, or by the
TDDGQO parameter of the PTPPKF object for TDD cells
– GFDD_Qmin: minimum threshold for Ec/No for UMTS FDD cell re-selection; the
user sets this value by the GFDDQMI parameter of the PTPPKF object
If the GPRS 3G Cell Reselection list includes UTRAN frequencies, the MS will, at least
every 5 second, update the value RLA_P for the serving cell and each of the at least 6
strongest non-serving GSM cells.
The MS will then reselect a suitable UTRAN cell if its measured RSCP value exceeds
the value of RLA_P for the serving cell and all of the suitable non-serving GSM cells by
the value XXX_GPRS_Qoffset for a period of 5 seconds and (only in case of FDD cells)
the UTRAN cells measured Ec/No value is equal or greater than the value FDD_Qmin.
If a cell reselection occurrs within the previous 15 seconds, XXX_GPRS_Qoffset is
increased by 5 dB.
If more than one UTRAN cell fulfills the above criteria, the MS selects the cell with the
highest RSCP value.
If the MS has reselected a GSM cell from an UTRAN one, cell reselection to UTRAN
does not occur within 5 seconds, if a suitable GSM cell can be found.
There is also a threshold by which the network indicates whether or not the measure-
ments for the cell reselection of the UMTS cells should be performed; the threshold indi-
cates if the signal level of the serving cell should be below or above it, in order to perform
UMTS cells measurements; the user sets this value by the QSRHPRI parameter of the
PTPPKF object.
FDDGQO, TDDGQO, GFDDQMI, and QSRHPRI parameters are broadcast on the
i PBCCH of the serving cell.

10.2.3 Handling of UMTS Neighboring Cells


Besides defining the re-selection criteria, the user must also define the UMTS neigh-
boring cells to be re-selected (obviously a UMTS cell is always considered an external
cell, i.e., a cell that does not belong to the BSC). To define a UMTS neighboring cell for

250 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

a specific BTS object instance, the user create an instance of the ADJC3G object
(subordinated to the BTS one).
For each BTS object instance the user can define up to 64 neighboring UMTS cells
i (ADJC3G object).
For each BTS object instance, the user can define up to 32 neighboring
GSM/GPRS/EGPRS cells (ADJC object) if there are no UMTS neighboring cells, and
up to 31 neighboring GSM/GPRS/EGPRS cells (ADJC object) if UMTS neighboring
cells are defined.

The TGTCELL parameter of the ADJC3G object contains a reference to the following:
• a TGTFDD object instance, in case of FDD neighboring cell; a TGTFDD managed
object instance contains all of the parameters that allow describing, in the BSC data-
base, the external UMTS FDD cell (the same principle as described in
"10.1.4.1 Handling of Neighboring Cells" is also used to manage external UMTS
cells).
The following are the more important parameters of the TGTFDD object:
– CELLGLID (C-ID cell identifier): identifies univocally the UMTS FDD cell in the
UMTS/GSM networks and it is composed by MCC (Mobile Country Code), MNC
(Mobile Network Code), LAC (Location Area Code) and CI (Cell Identifier)
– FDDARFCN: defines the frequency of the cell
– RNCID: identifies the RNC
– FDDSCRMC: defines the scrambling code
– FDDDIV: indicates if diversity is applied for the cell
• a TGTTDD object instance, in case of TDD neighboring cell; a TGTTDD managed
object instance contains all the parameters that allow describing, in the BSC data-
base, the external UMTS TDD cell.
The following are the more important parameters of the TGTTDD object:
– CELLGLID (C-ID cell identifier): it identifies univocally the UMTS TDD cell in the
UMTS/GSM networks and it is composed by MCC (Mobile Country Code), MNC
(Mobile Network Code), LAC (Location Area Code) and CI (Cell Identifier);
– TDDARFCN: it defines the frequency of the cell;
– RNCID: it identifies the RNC;
– BNDWIDTDD: it defines the bandwidth used for TDD;
– TDDDIV: it indicates if diversity is applied for the cell.
Therefore, before creating the ADJC3G object related to an UMTS neighboring cell of a
specific BTS, the user must already have created either the TGTFDD or the TGTTDD
object defining the UMTS cell.
In this way, different BTS objects, that have the same UMTS cell as neighboring cell, will
indicate the same TGTFDD (or the same TGTTDD) object instance in the adjacent rela-
tionship defined by the subordinate ADJC3G object instance.
EXAMPLE: if the TGTFDD:0 instance has been created to define a UMTS cell in the
BSC database, this UMTS cell can be defined as adjacent to both the BTS:1 and BTS:5
cells in the following way:
– if, for example, the ADJC3G:4 instance of the BTS:1 object represents the neigh-
boring relationship towards the UMTS cell defined by the TGTFDD:0 instance, the
user sets the TGTCELL attribute equal to TGTFDD:0;
– if, for example, the ADJC3G:2 instance of the BTS:5 object represents the neigh-
boring relationship towards the UMTS cell defined by the TGTFDD:0 instance, the
user sets the TGTCELL attribute equal to TGTFDD:0.

A30808-X3247-L24-5-7618 251
GPRS/EGPRS Global Description Information
System

10.3 Network Controlled Cell Reselection and Traffic Control


Management
As described in "10.1 Cell Selection and Re-selection", the cell reselection algorithm is
executed normally by the mobile station. Every MS in packet idle mode and in packet
transfer mode measures received signals from both the serving cell and neighboring
cells, and performs autonomously cell reselection.
The Network Controlled Cell Reselection is another available cell reselection method:
the network may request the measurement reports from the MSs and control their cell
reselection.
Therefore, if the user enables this feature, the network can ask the mobile to transmit
the carrier level of both serving and adjacent cells through packet measurement reports;
depending on these reported values, the network can transfer a mobile station to a
another cell, which is better from a radio condition point-of-view. This algorithm is called
Radio Link Network Controlled Cell Reselection because the network cell reselection
brings mobile stations to another cell that is better from the radio condition point-of-view.
However there is another topic that must be considered: the BSC allocates PDCHs as
long as there are available resources in a given cell. This might lead to congestion,
although traffic capacity might be available in neighboring cells. Therefore, if the user
enables the Traffic Control feature, the network may redistribute MSs among cells to
satisfy the maximum number of service requests, i.e., the Traffic Control Network
Controlled Cell Reselection can be executed.
The Traffic Control network controlled cell reselection guarantees the optimum usage of
resources, i.e., the better traffic distribution among the available channels in all of the
available cells. So, even if the handover functionality is not foreseen for GPRS/EGPRS
services, the functionality of the traffic control network controlled cell re-selection has
the same purpose of the handover due to traffic, i.e., the distribution of MSs among cells
according to network criteria.
The following must be clear:
a) if the user enables only the network controlled cell reselection feature, only the
Radio Link controlled cell reselection is enabled;
b) if the user wants to enable the Traffic Control controlled cell reselection, he/she must
enable, besides the network controlled cell reselection, the traffic control strategy.
The topics below are described in the following paragraphs:
– 10.3.1 describes how the network controlled cell reselection works
– 10.3.1.1 describes some notes about packet measurement reports
– 10.3.1.2 describes the Radio Link network controlled cell reselection algorithm
– 10.3.1 and 10.3.2.1 describe the Traffic Control strategy and the related traffic
control network controlled cell reselection algorithm

10.3.1 Network Controlled Cell Reselection


With Network controlled cell reselection, the network may request measurement reports
from the MSs and control their cell re-selection.
The NTWCOR (NETWORK_CONTROL_ORDER) parameter indicates if and how the
network controls the reselection process. The meaning of different values of the
NTWCOR parameter is specified as follows:

252 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

• NC0: normal MS control; the MS performs autonomous cell re-selection as


described in "10.1 Cell Selection and Re-selection"
• NC1: MS control with measurement reports; the MS sends measurement reports to
the network, but it performs autonomous cell re-selection
• NC2: network control; the MS sends measurement reports to the network, and it
does not perform autonomous cell re-selection
The NETWORK_CONTROL_ORDER parameter is broadcast from the network to all
MS in the cell, by PSI5 on PBCCH or SI13 on BCCH. Alternatively, the network can use
the Packet_Measurement_Order or Packet_Cell_Change_Order messages on PACCH
to address a particular MS.
To enable the Network Controlled Cell Reselection feature, the user must set the NCRE-
SELFLAG parameter, belonging to the BSC object, at ENABLE.
Remember that when the user set the NCRESELFLAG parameter at ENABLE, only the
i radio link controlled cell reselection is enabled.

When the feature is enabled, the network can ask the mobile (by setting the NTWCOR
parameter for that mobile, see below) to transmit the carrier level of both serving and
adjacent cells; then the MS sends measurement reports periodically. The period is
defined by two attributes:
– NTWCREPPIDL (NC_REPORTING_PERIOD_I) for MS in idle mode
– NTWCREPPTR (NC_REPORTING_PERIOD_T) for MS in transfer mode
Regarding measurement reports, there is also the NTWCNDRXP parameter that
i defines the minimum time the mobile station will stay in non-DRX mode after a measure-
ment report has been sent with the mobile in packet idle mode; however this parameter
is not used, since MSs in packet idle mode do not send measurement reports.

GPRS and EGPRS mobiles in packet idle mode always work in NC0 mode, otherwise
the network would have to manage the packet measurement reports and associated
access requests needed by mobiles to transmit periodically packet measurement
reports. In fact, taking into account that the longest period of transmission of packet
measurement report for mobile in packet idle mode is about 60 seconds, at least 60
channel requests per mobile per hour must be considered only for measurement report
transmission; this would hardly increase PCU real time requirements. In addition, there
are impacts even on battery power safe.
Consequently, NC2 will be used only for mobiles in packet transfer mode, which will then
submit measurement reports with the reporting period defined by NTWCREPPTR.
Therefore, if the network controlled cell reselection is enabled (NCRESELFLAG set at
ENABLE) things work in the following way.
The NTWCOR broadcast value (PSI5 on PBCCH or SI13 on BCCH) is always NCO, so
every mobile station in packet idle mode does not transmit any packet measurement
report to the BTS.
When a GPRS/EGPRS mobile station is involved in a TBF (uplink or downlink), the BSC
modifies the NTWCOR mode value, from NC0 to NC2, by the Packet Measurement
Order message, transmitted to that single mobile on PACCH. This message also carries
the NTWCREPPTR parameter, which overwrites the correspondent value optionally
broadcasted by PSI5 or SI13. In the Packet Measurement Order message, NTWCN-
DRXP and NTWCREPPIDL have no significant value (MS transmits packet measure-
ment report only in packet transfer mode).

A30808-X3247-L24-5-7618 253
GPRS/EGPRS Global Description Information
System

After this change, mobile working in NC2 control mode periodically transmits Packet
Measurement Report messages to the BSC:
• if MS is involved in uplink TBF, it uses an USF scheduled block
• if Ms is involved in downlink TBF, it uses a RRBP assigned block
If needed conditions are verified (see "10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm" and "10.3.2.1 Network Controlled Cell Reselection Algorithm for
Traffic Control Strategy") the BSC may transfer the MS to another cell by a Packet Cell
Change Order message; this message contains the following:
– characteristics of the new cell that are necessary to identify it (i.e., BSIC + BCCH
frequency)
– network controlled measurement parameters valid for the mobile station in the new
cell (e.g., NTWCREPPTR)
– IMMEDIATE_REL parameter
Upon receipt of the Packet Cell Change Order message the mobile station starts timer
T3174. When a network controlled cell reselection is made, the mobile station will act
upon the IMMEDIATE_REL value which has been received in the Packet Cell Change
Order. If required, it will immediately abort any TBF in progress by immediately ceasing
to decode the downlink and transmit on the uplink, stopping all RLC/MAC timers except
for timers related to measurement reporting, otherwise the mobile station may continue
its operation in the old serving cell until TBF end. The mobile station will then switch to
the identified new cell and will obey the relevant RLC/MAC procedures on this new cell.
The mobile station regards the procedure as completed when it has received a
successful response to its access request on the new cell.
If timer T3174 expires before a response to the access request message has been
received on the new cell, or if an IMMEDIATE ASSIGNMENT REJECT or PACKET
ACCESS REJECT message is received from the new cell, or if the contention resolution
procedure fails on the new cell, then the mobile station will start the T3176 timer and
return to the old cell.
If the mobile station was in uplink packet transfer mode or in a simultaneous uplink and
downlink packet transfer mode before the cell change, the mobile station will establish
a new uplink TBF and send the PACKET CELL CHANGE FAILURE message on this
TBF. The mobile station will then resume its uplink transfer on this TBF. When the
mobile station has sent a PACKET CELL CHANGE FAILURE message, timer T3176 will
be stopped. If T3176 expires and the mobile station was previously in an uplink packet
transfer mode or in a simultaneous uplink and downlink packet transfer mode on the old
cell, the mobile station will perform the abnormal release with random access.
If the mobile station was previously in a downlink packet transfer mode only on the old
cell, the mobile station will perform an abnormal release with return to CCCH or PCCCH.
On the mobile station side, if the Packet Cell Change Order message instructs the
mobile station to use a frequency that it is not capable of using, then the mobile station
will return a PACKET CELL CHANGE FAILURE message with cause "frequency not
implemented". If the Packet Cell Change Order message is received by the mobile while
a circuit switched connection is on-going, then the mobile station will return a PACKET
CELL CHANGE FAILURE message with the cause "on-going CS connection".
When a network controlled cell reselection occurs (ordered by the BSC), the BSS will
signal this exception condition to a SGSN by sending a RADIO-STATUS PDU (Radio
Cause value: cell reselection ordered). It will contain a reference to the MS, (either TLLI
or TMSI or IMSI).

254 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

This condition indicates that the SGSN should wait for a cell update or a routing area
update before resuming the transmission of LLC PDU to the BSS.
When the MS changes the cell, it starts a cell update procedure or a routing area update
procedure towards the SGSN.
After this procedure, the SGSN transmits the FLUSH_LL message towards the BSC
indicating the new cell where the MS is entered.
The BSC uses this indication to route the queued RLC blocks related to that MS; if the
cell belongs to a different PPXU, the queued RLC blocks are discarded. Then the BSC
transmits the FLUSH_LL_ACK message to the SGSN, indicating if re-route or discard
is made. It is responsibility of the higher layer protocols in the SGSN to cope with
discarded LLC frames. If new cell belongs to another SGSN, an inter_SGSN routing
area update is required before the TBF starts in the new cell (Fig. 10.2 shows this
procedure).
Before ending TBF, the BSC changes the network control mode to NC0, so when this
mobile station enters the packet idle mode, it no longer transfers packet measurement
reports.

MS BSC SGSN
Packet Cell Change Order

Start T3174
Radio Status

Cell Change

Packet Channel Request

Packet Uplink Assignment


Stop T3174
RLC Block

RLC Block

LLC (Cell Update)

Flush LL

Fig. 10.2 Network Controlled Cell Reselection Procedure

10.3.1.1 Measurement Reporting


After each reporting period defined by NTWCREPPTR, the MS sends a measurement
report to the BSS.
The MS will then discard any previous measurement report, that it has not been able to
send. The Packet Measurement Report contains:

A30808-X3247-L24-5-7618 255
GPRS/EGPRS Global Description Information
System

– RXLEV for the serving cell


– received signal level for the non-serving cells
For normal measurement reporting, carriers will be reported if they are among the 6
strongest carriers and BSIC is successfully decoded and allowed i.e., either equal to the
BSIC of the list or with allowed NCC part of BSIC. In the latter case, which applies for
BA(BCCH) where no BSIC is given, the decoded BSIC will be included in the report.
In the case of a multiband MS, the MS will report the number of the strongest BCCH
carriers in each band as indicated by the GNMULBAC parameter, broadcast on
PBCCH, or if PBCCH does not exist, on BCCH.
For multi-RAT MS, the MS will report the number of valid cells in each other radio access
technology as indicated by specific parameters: GFDDMURREP (GPRS FDD
MULTIRAT REPORTING) and GTDDMURREP (GPRS TDD MULTIRAT REPORTING)
parameters define the number of valid UMTS neighbor cells (FDD and TDD) which will
be reported by the MS/UE. The remaining positions in the measurement report will be
used for reporting of GSM cells. If remaining positions still exist, these will be used to
report the next valid UMTS cells. In this case, the received signal level is replaced by the
relevant measurement quantity.
The UMTS FDD measurement quantity Ec/No is not a suitable parameter for a compar-
! ison with the GSM received level because the Ec/No is a quality parameter and not a
received level parameter.
Therefore, the GFDDREPQTY parameter (FDD_REP_QUANT) is introduced in order to
tell the GPRS/EGPRS attached MS/UE whether to report the RSCP value
(GFDDREPQTY=RSCP) or the Ec/No one (GFDDREPQTY=EC_NO) to the BTS.

Network Controlled Cell Reselection towards the UMTS network is not supported in
i BR7.0.

10.3.1.2 Radio Link Network Controlled Cell Reselection Algorithm


When the network controlled cell reselection is enabled (i.e., the NCRESELFLAG
parameter is set at ENABLE) the radio link controlled cell reselection algorithm is
executed by the BSC.
The algorithm works independently if the GPRS/EGPRS control strategy (see
i "10.3.2 GPRS/EGPRS Traffic Control Strategy") has been enabled by the operator.

When the radio scenario of the mobile station is degraded, the BSC chooses a better
neighboring cell and commands that mobile to move on this new cell.
According to what was described in "10.3.1 Network Controlled Cell Reselection", the
mobile station sends measurement reports to the BSC; when the BSC receives a packet
measurement report from a mobile, the following values are calculated:
– the C1 value for the serving cell [C1(s)]
– and the C1 value for each adjacent cell [C1(n)] reported in the packet measurement
report
The C1 value for both serving and neighboring cells is calculated with the following
criteria:

C1 = RLA_P – GPRS_RXLEV_ACCESS_MIN – Max( 0,


GPRS_MS_TXPWR_MAX_CCH – P)

256 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Where:

- P is the Power Class of the MS


- GPRS_RXLEV_ACCESS_MIN is the minimum allowed received level to access a
cell; the user can define this value by the GRXLAMI parameter (for the serving cell this
parameter is set in the PTPPKF object, for neighboring cells it is set in the TGTPTPPKF
one)
- GPRS_MS_TXPWR_MAX_CCH is the maximum power that the MS can use to
access the cell; the user can define this value by the GMSTXPMAC parameter (for the
serving cell this parameter is set in the PTPPKF object, for neighboring cells it is set in
the TGTPTPPKF one).

If C1(s) < NCC1TH, the mobile must be moved to another cell, to avoide losing it.
The PKTMEASREPCNT parameter specifies how many consecutive measurements of
i the BCCH carrier of the serving cell, under the NCC1TH threshold, are necessary to
order a cell change.

First of all, among the neighboring cells reported in the packet measurement report, only
those for which C1 (n) > NCC1THADJC are selected.
Then, according to the value of the NCSARA attribute, the selected cells are ordered
according to different priorities.
If NCSARA = TRUE, the adjacent cell is searched, before, among cells belonging to the
same routing area of serving cell; therefore the following priorities are used to order
cells:
1. target cell with the same Routing Area on the same PPXU/PPCU
2. target cell with different Routing Area but on the same PPXU/PPCU
3. target cell with the same Routing Area on different PPXU/PPCU and same BSC
4. target cell with different Routing Area on different PPXU/PPCU and same BSC
5. target cell on different PPXU/PPCU and different BSC
If NCSARA = FALSE, adjacent cells of the same routing area have no priority compared
to adjacent cells of other routing areas; therefore the following priorities are used to
order cells:
1. target cell with the same RA or target cell with different RA, but on the same
PPXU/PPCU
2. target cell on different PPUX/PPCU and same BSC
3. target cell on different PPUX/PPCU and different BSC
Among neighboring cells with the same priority, the cell with the highest C32(n) value is
chosen. The C32(n) value is calculated as follows:

there are two cases:


If T <= NC_GPRS_PENALTY_TIME
C32(n) = C1(n) + NC_GPRS_RESELECT_OFFSET(n) –
NC_GPRS_TEMPORARY_OFFSET(n)

If T > NC_GPRS_PENALTY_TIME
C32(n) = C1(n) + NC_GPRS_RESELECT_OFFSET(n)
Where:
• NC_GPRS_RESELECT_OFFSET(n) is a positive offset that increases the priority
of cell in the list of the strongest neighbor cells. The user sets a

A30808-X3247-L24-5-7618 257
GPRS/EGPRS Global Description Information
System

NC_GPRS_RESELECT_OFFSET(n) value for every adjacent relationship, by


NCGRESOFF parameter of the ADJC object;
• NC_GPRS_TEMPORARY_OFFSET(n) applies a negative offset to C32 for the
duration of NC_GPRS_PENALTY_TIME(n) after the timer T has started for that cell.
The T timer is started for each cell in the list of the 6 strongest neighboring cells, as
soon as it is placed on the list. T is reset to 0 if the cell is removed from the list.
NC_GPRS_PENALTY_TIME is the duration for which
NC_GPRS_TEMPORARY_OFFSET applies.
The user sets a NC_GPRS_TEMPORARY_OFFSET(n) value and a
NC_GPRS_PENALTY_TIME(n) value for every adjacent relationship, by NCGTEM-
POFF and NCGPENTIME parameters of the ADJC object.
When NCSARA is set at FALSE, a hysteresis value is subtracted from the C32 value for
the neighbor cells. The hysteresis value can be set by the user via the NCRARESH
parameter belonging to the PTPPKF object. NCRARESH must be set at DB00 (default
value) when NCSARA is set at TRUE.
Moreover, in order to prevent “ping_pong” effect due to questionable MS behaviour
during Network Controlled Cell Reselection, the TRFPSCTRL parameter in the object
PTPPKF is used to avoid too frequent cell reselection of the same adjacent cell.To this
end, BSC doesn’t order to MS to move again into same adjacent target cell where a
NCCR failed,in spite of good radio link scenario ,until timer TRFPSCTRL is expired and
the following condition is satisfyied:
– STGTTLLIINF > TRFPSCTRL
If the STGTTLLIINF parameter set to NULL, this means that no TBF temporary data is
i stored and therefore the ping pong NCCR cannot be avoided.

10.3.2 GPRS/EGPRS Traffic Control Strategy


The GPRS/EGPRS Traffic Control Strategy feature makes it possible to control the traffic
distribution among cells belonging to the same PCU; the feature is based on the
Network Controlled Cell Reselection one (see 10.3.1) and on appropriate traffic thresh-
olds set in each cell.
In fact, the Network Controlled Cell Re-Selection feature introduces the management of
measurements related to the neighboring cells reported by the GPRS/EGPRS MS, but
it does not specify any traffic control strategy on how to run this available information
(only radio link conditions are taken into account); the GPRS/EGPRS Traffic Control
Strategy feature exploits this information to distribute the traffic among all available
network resources.
To enable the Traffic Control Strategy feature at BSC level, the user must set the TRFPS
(trafficPs) parameter to TRUE
The feature can be enabled only if Network Controlled Cell Reselection feature
i (see 10.3.1) is already enabled.

When TRFPS is set to TRUE, the traffic control algorithm is applied (see
"10.3.2.1 Network Controlled Cell Reselection Algorithm for Traffic Control Strategy").
The feature goal is to spread the cell traffic on more than one cell, that is to move MSs
inside a high traffic cell towards available resources in neighboring cells.
Traffic control algorithm is applied only to cells belonging to the same PCU, because
i every PCU knows only its own traffic.

258 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Traffic control algorithm performs an evaluation of the radio resource occupation into
each cell, based on the number of channels configured, and in service available for
GPRS/EGPRS, and the type of strategy set by the operator.

10.3.2.1 Network Controlled Cell Reselection Algorithm for Traffic Control


Strategy
When the Traffic control strategy is enabled, every TBF activation is checked if, for a
specific cell, the radio resource occupation has reached or exceeded a threshold,
defined by the CRESELTRHSOUT parameter.
The radio link criteria, described in "10.3.1.2 Radio Link Network Controlled Cell Rese-
i lection Algorithm" is maintained also when Traffic Control Strategy is enabled.

In the positive case, the algorithm looks for MSs candidates to be forced to a cell rese-
lection. The mobile(s) to move are chosen among those in packet transfer mode,
applying the following criteria for the choice of target cell.
First of all, among the neighboring cells reported in the packet measurement report, only
those for which C1 (n) > NCC1THADJC are selected.
Then, according to the value of the NCSARA attribute, the selected cells are ordered
according to different priorities.
If NCSARA = TRUE, the adjacent cell is searched, before, among cells belonging to the
same routing area of serving cell; therefore the following priorities are used to order
cells:
1. target cell with the same Routing Area on the same PPXU/PPCU
2. target cell with different Routing Area, but on the same PPXU/PPCU
If NCSARA = FALSE, adjacent cells of the same routing area have no priority compared
to adjacent cells of other routing areas; therefore only priority level exists, i.e., target cell
with the same RA or target cell with different RA, but on the same PPXU/PPCU
Among neighboring cells with the same priority, the cell with the highest C32(n) value is
chosen. The C32(n) value is calculated as follows:

there are two cases:


If T <= NC_GPRS_PENALTY_TIME
C32(n) = C1(n) + NC_GPRS_RESELECT_OFFSET(n) –
NC_GPRS_TEMPORARY_OFFSET(n)

If T > NC_GPRS_PENALTY_TIME
C32(n) = C1(n) + NC_GPRS_RESELECT_OFFSET(n)
Where:
• NC_GPRS_RESELECT_OFFSET(n) is a positive offset that increases the priority
of the cell in the list of the strongest neighbor cells. The user sets a
NC_GPRS_RESELECT_OFFSET(n) value for every adjacent relationship, by
NCGRESOFF parameter of the ADJC object;

A30808-X3247-L24-5-7618 259
GPRS/EGPRS Global Description Information
System

• NC_GPRS_TEMPORARY_OFFSET(n) applies a negative offset to C32 for the


duration of NC_GPRS_PENALTY_TIME(n) after the timer T has started for that cell.
The T timer is started for each cell in the list of the 6 strongest neighboring cells, as
soon as it is placed on the list. T is reset to 0 if the cell is removed from the list.
NC_GPRS_PENALTY_TIME is the duration for which
NC_GPRS_TEMPORARY_OFFSET applies.
The user sets a NC_GPRS_TEMPORARY_OFFSET(n) value and a
NC_GPRS_PENALTY_TIME(n) value for every adjacent relationship, by NCGTEM-
POFF and NCGPENTIME parameters of the ADJC object.
When NCSARA is set to FALSE, a hysteresis value is subtracted from the C32 value for
the neighbor cells. The hysteresis value can be set by the user via the NCRARESH
parameter, belonging to the PTPPKF object. NCRARESH must be set at DB00 (default
value) when NCSARA is set at TRUE.
The algorithm also looks for a possible candidate cell into which to move a MS. A cell
can be a candidate for this procedure only if:
– it belongs to the same PCU of the serving cell
– it is adjacent to the origin cell (i.e., the relevant ADJC object already exists)
– it is not barred
– it supports the GPRS service and its resource occupation is under a threshold. This
threshold can be set by the user by means of the CRESELTRSHINP parameter.
Then, the number of MSs to be forced in reselection is determined taking as many MSs
as the radio resources that must be released, in order to put the traffic load under the
NCTRFPSCTH parameter.
The network sends to each concerned MS a PACKET CELL CHANGE ORDER
message with the indication of the new cell where the MS must perform the cell reselec-
tion.

Details About the Calculation of the “Allocated Resources”.


As described, when the traffic control strategy is enabled, the following statements are
valid:
– every TBF activation is checked if the radio resource occupation has reached or
exceeded a threshold, defined by the CRESELTRHSOUT parameter, for a specific
cell
– the algorithm looks for a possible candidate cell into which to move a MS; the
resource occupation of the candidate cell must be under a threshold. This threshold
can be set by the user with the CRESELTRSHINP parameter
– the number of MSs to be forced in reselection is determined taking as many MSs as
the radio resources that have to be released, in order to put the traffic load under the
NCTRFPSCTH parameter.
The traffic control strategy is active when vertical allocation is present and it is based on
the calculation of the “allocated resources”, according to the following rules.
The calculation is applied both to uplink and downlink TBFs: it considers the number of
allocated timeslots, the GMANMSAL attribute value and weight factors, which take into
account the average number of timeslots allocated for each TBF (uplink or downlink).
The algorithm calculates the available resources both in the uplink and downlink direc-
tions:

260 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

AVAL_TBF_DL= [N_GPRS_allocated_ts * GMANMSAL_DL] / aver_ass_ts_dl

AVAL_TBF_UL= [N_GPRS_allocated_ts * GMANMSAL_UL] / aver_ass_ts_ul

where:
– N_GPRS_allocated_ts is the number of allocated time slots when vertical allocation
is present;
– aver_ass_ts_ul and aver_ass_ts_dl are weighted factors which consider number of
time slots assigned to the uplink and downlink TBF on average.
The system calculates the “traffic percentage“ both for uplink and for downlink direc-
tions, each time a TBF or a PDCH is added or removed:

PercTrfUL = TBF_UL / AVAL_TBF_UL


PercTrfDL = TBF_DL / AVAL_TBF_DL

where TBF_UL and TBF_DL indicate the number of currently opened uplink and down-
link TBFs, taking into account the related weight factors aver_ass_ts_ul and
aver_ass_ts_dl.
If TRFPS is set to TRUE and if the vertical allocation is used, the system checks the
following conditions:
1. (PercTrfUL > CRESELTRHSOUT) OR (PercTrfDL > CRESELTRHSOUT);
2. (PercTrfUL < NCTRFPSCTH) AND (PercTrfDL < NCTRFPSCTH);
3. (PercTrfUL_adjc > CRESELTRSHINP ) OR (PercTrfDL_adjc > CRESELTRSHINP).
Then:
a) when condition 1) is satisfied, the system moves a mobile station from the serving
cell to a suitable adjacent cell; this process of moving suitable mobile stations
continues until condition 2) is reached;
b) when condition 2) is satisfied, the system stops moving mobiles to the adjacent cell
for traffic reason;
c) when, for an adjacent cell, condition 3) is verified, this adjacent cell is no longer suit-
able to accept mobile from a congested cell.
The process is stopped when a transition from the vertical allocation to the horizontal
allocation is executed.

10.4 Power Control


The objective of the power control feature is to adapt the transmitted power of the MS,
as well as of the BTS, to the reception conditions. The following are the two advantages
of the power control algorithm:
1. reduction of the power consumption of the MS’s batteries
2. reduction of the interference which is experienced by both co-channel and neigh-
boring channel users.
For the uplink, the MS follows a flexible power control algorithm, which the network can
optimize through a set of parameters. The algorithm can be used for both of the following
power control methods:

A30808-X3247-L24-5-7618 261
GPRS/EGPRS Global Description Information
System

– open loop power control: the MS output power is based on the received signal
strength at the MS side, assuming the same path loss in uplink and downlink direc-
tions
– closed loop power control: the MS output power is commanded by the network
based on signal measurements made in the BTS.
In BR7.0 only open loop power control is supported.
For the downlink, the power control is performed in the BTS. Therefore, there is no need
to specify the actual algorithm, but information about the downlink performance is
needed. Therefore, the MSs must transfer Channel Quality reports to the BTS.
In BR7.0 downlink power control is not supported: power control is a mandatory
feature for the MS, while it is optional for the network.

10.4.1 Power Control Algorithm


Even if the BSS functionality is not directly involved, this paragraph provides a brief
i description of the power control algorithm implemented in the mobile stations.

The RF output power, Pch, to be employed by the MS on each individual uplink PDCH
is given by the following formula:

Pch= min (GAMMA0 - GAMMAch - ALPHA * (C+48), Pmax) (1)

where:

GAMMAch: is a MS and channel specific power control parameter, sent to the MS


in control messages such as IMMEDIATE ASSIGNMENT, PACKET
UPLINK/DOWNLINK ASSIGNMENT, PACKET UPLINK ACK/NACK.
The operator can set this value using the GAM parameter of the
PTPPKF object.
GAMMA0: = 39 dBm for GSM900
= 36 dBm for GSM1800.
ALPHA: is the ALPHA system parameter, which is broadcast on the
BCCH/PBCCH or optionally sent to MS in a control message.
C: is the normalized received signal level at the MS side
Pmax: is the maximum allowed output power in the cell. It is equal to:
- GMSTXPMAC if PBCCH is configured in the serving cell
- MSTXPMAXCH (i.e., the analogous GSM parameter) otherwise
Power levels are expressed in dBm.

When the MS receives either a new GAM or ALPHA value, the MS will use the new
value to update Pch according to equation (1).
The MS uses the same output power on all four bursts within one radio block.
When accessing a cell on PRACH or RACH (random access) and before receiving the
first power control parameters during packet transfer on PDCH, the MS will use the
output power defined by Pmax.
If a calculated output power is not supported by the MS, the MS uses the supported
output power that is closest to the calculated output power.

262 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

10.4.2 Measurement at the MS Side


A procedure is implemented in the MS to monitor periodically downlink received signal
level and quality from its serving cell.
To calculate the Pch value according to equation (1), the MS must derive the C value.
Two different methods are used to estimate the C value, according to the MS state (i.e.,
packet idle mode or packet transfer mode).

10.4.2.1 Packet Idle Mode: Measurements for Power Control


In packet idle mode, the MS periodically measures the received signal level of the
PCCCH or, if PCCCH is not configured, of the BCCH. The MS measures the received
signal level of each paging block monitored by the MS according to its current DRX
mode and its paging group (see "9.8.3.2 Discontinuous Reception").
The normalized C value for each radio block is calculated as follows:

Cblock(n)= SSblock(n) + Pb (2)

where:

SSblock(n): is the mean value of the received signal levels on the four normal
bursts that compose the block;
Pb: is the BTS output power reduction (relative to the output power used
on BCCH) used on the channel on which the measurements are
performed; it corresponds to the PRPBCCH parameter. For PCCCH,
Pb is broadcast on PBCCH. For BCCH, Pb = 0 (not broadcast).

Finally, the Cblock(n) values are filtered with a running average filter:

C(n)= (1- a)* C(n- 1) + a * Cblock(n) C(0)=0

where a is the forgetting factor:

a = 1/MIN(n, MAX (5, TAVG_W/TDRX)

TDRX: is DRX period for the MS (see "9.8.3.2 Discontinuous Reception");


TAVG_W: is the TAVGW parameter and indicates the signal strength filter period
for power control in packet idle mode. It is broadcast on PBCCH or, if
PBCCH does not exist, on BCCH.
n n is the iteration index. The filter will be restarted with n=1 for the first
sample every time a new cell is selected. Otherwise, when entering
packet idle mode, the filter will continue from the n and C(n) values
obtained during packet transfer mode. The filter will also continue from
its previous state if TDRX is changed.

A30808-X3247-L24-5-7618 263
GPRS/EGPRS Global Description Information
System

The current C(n) value is used in formula (1) to calculate the output power when the MS
transfers its first radio block.

10.4.2.2 Packet Transfer Mode: Measurements for Power Control


In packet transfer mode, the MS uses the same received signal level measurements on
the BCCH carrier of the serving cell as made for cell reselection (see "10.4.2.2 Packet
Transfer Mode: Measurements for Power Control"). The measurements are filtered with
a running average filter:

C(n)= (1- b)* C(n- 1) + b * SS(n)

where b is the forgetting factor:

b = 1/(6 * TAVG_T)

and:

SSn: is the received signal level of the measurement samples;


TAVG_T: is the TAVGT parameter and indicates the signal strength filter period
for power control in packet transfer mode. It is broadcast on PBCCH
or, if PBCCH does not exist, on BCCH.
n n is the iteration index; when entering packet transfer mode, the filter
will continue from the n and C(n) values obtained during packet idle
mode.

If indicated by the PCMECH (PC_MEAS_CHAN) parameter, the MS will instead


measure the received signal level of each radio block on one of the PDCH monitored by
the MS for PACCH. For each downlink radio block a Cblock(n) value is derived
according to formula (2) (if PBCCH does not exist, Pb = 0).
Finally, the Cblock(n) values are filtered with a running average filter:

C(n)= (1- c)* C(n- 1) + c * Cblock(n)

where c is the forgetting factor:

b = 1/(12 * TAVG_T)

and:

TAVG_T: is the TAVGT parameter and indicates the signal strength filter period
for power control in packet transfer mode. It is broadcast on PBCCH
or, if PBCCH does not exist, on BCCH.
n n is the iteration index; when entering packet transfer mode, the filter
will continue from the n and C(n) values obtained during packet idle
mode.

264 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Once the current C(n) value has been obtained, this value is used to update formula (1).
Each time a new C(n) value is obtained or whenever the MS applies new GAM or
ALPHA values, the Pch value is updated.

10.4.2.3 Derivation of Channel Quality Reports


The channel quality is measured as the interference signal level during the idle frames
of the multiframe, when the serving cell is not transmitting.
In packet transfer mode, the MS measures the interference signal strength of all eight
channels (slots) on the same carrier as the assigned PDCHs. The MS will make these
measurements during the search frames and PTCCH frames. The measured interfer-
ence will be averaged in a running average filter. For each channel, the MS will perform
at least NAVGI measurements before valid running average filter values can be deter-
mined.
NAVGI is broadcast on the PBCCH or, if the PCCH does not exists, on the BCCH.
In packet idle mode, the MS will measure the interference signal strength on certain
channelsthat are indicated on the PBCCH or if the PBCCH does not exist, on the BCCH.
The interference measurements will be made and averaged in the same way as for
packet transfer mode.

10.4.3 BTS Output Power


The BTS uses constant power on those PDCH radio blocks that contain PBCCH or
which may contain PCCCH. This power may be lower than the output power used on
BCCH. The difference will be broadcast on PBCCH by means of the PRPBCCH param-
eter.

10.5 Link Adaptation


As previously described (see 4.2), GPRS offers four different coding schemes, instead,
for EGPRS nine different modulation and coding schemes are defined.
In both cases (GPRS and EGPRS) lowest coding schemes (e.g., CS1 and MCS1
respectively) show good performances in poor radio conditions; on the other hand, only
the highest coding schemes (e.g., CS4 and MCS9 respectively) can provide high data
throughput in good radio environments.
Therefore, an algorithm is needed to dynamically select the coding scheme that
behaves better in a specific radio condition. The dynamic selection of the coding
scheme, to suit radio link quality, is referred to as Link Adaptation.
Therefore, the basic idea is to dynamically select the coding scheme that allows the
highest throughput according to the present radio conditions. Then the problem is to find
the switching points that allow changing from one coding scheme to another.
Link Adaptation can be enabled, for both GPRS and EGPRS services, using the
ELKADPT parameter of the PTPPKF object. Link Adaptation is enabled in both uplink
and downlink directions at the same time.
The advantage to switching to a more robust coding scheme can be seen in Fig. 10.3,
taking into account GPRS CS2 and CS1 coding schemes.

A30808-X3247-L24-5-7618 265
GPRS/EGPRS Global Description Information
System

Gross CS2
Throughput
[kbit/s]

CS1

C/I [dB]

Fig. 10.3 CS1 and CS2 Throughput Depending on C/I (dB)

Assuming that the C/I ratio is better (i.e., higher direction towards '+') than the value
denoted with '='. In this case the use of the higher coding scheme (i.e., CS2) results in
an improved gross throughput compared to the use of CS1.
The situation changes, if the C/I becomes lower than '=' (direction towards '-'), according
to the propagation conditions. In this case, the use of CS2 results in a lower gross
throughput than with CS1. This due to the necessity to re-send many blocks because
they could not be received without errors the first time. In that situation, not only the
gross throughput is lower than possible (i.e., if CS1 had been used) but also the delay
increases. In other words: if conditions get worse, then a switch to the more robust
coding scheme improves gross throughput and reduces delay. On the other hand, if
propagation conditions improve, a switch to a higher coding scheme results in a better
gross throughput.
In general, C/I values are difficult to estimate in a real network, so another mechanism
is followed.
The triggering of the switch does not use separate measurements of channel quality, but
it is executed by analyzing the number of blocks to be repeated (not acknowledged
blocks) versus the number of transmitted blocks in total (i.e., the sum of the acknowl-
edged blocks and the unacknowledged one). Therefore to fix the switching points, the
NACK/(ACK+NACK) ratio (Block Erasure Ratio - BLER) is used; link adaptation is then
based on BLER measurements (indirect measures of the radio quality). The switching
points between coding schemes, to be used in link adaptation, are then defined in terms
of BLER thresholds (see 10.5.1 and 10.5.2.).
Since switching points depend on the actual RF scenario, it is impossible to calculate
such optimal values for each particular scenario. Upgrade switching points and down-
grade switching points are then stored in pre-calculated matrix tables, one for each
possible RF environment (these matrix tables cannot be set by O&M). By O&M, it is then
possible to select the suitable matrix table, containing all of the ideal switching points
(downgrade/upgrade switching points from/to all coding schemes) for the particular RF
scenario, by selecting the right radio environment.
The RAENV parameter allows the operator to specify the radio environment. As previ-
ously described, according to the chosen radio environment certain matrix tables are

266 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

selected (specific either for GPRS or for EGPRS) to define the BLER thresholds of the
switching points. The parameter can assume two values:
– LOWDIV (lowDiversity): it means that, for the MS, radio conditions can change
slowly, for example because Frequency Hopping is disabled and the cell is charac-
terized by low user mobility (e.g., because MS have a speed less than 50 Km/h or
because the cell is a small one);
– HIGHDIV (highDiversity): it means that, for the MS, radio conditions can change fast,
for example because Frequency Hopping is enabled.
Even if there are some common parameters to manage link adaptation, some differ-
ences exist between GPRS and EGPRS handling.
Besides in the current release an improvement of the Link Adaptation (LA)
GPRS/EGPRS algorithm’s thresholds has been implemented. The GPRS/EGPRS Link
Adaptation thresholds have been already optimized due to the improved RTT. Starting
from this first improvement the Threshold Table has been optimized again. For Uplink
and Downlink directions the Throughput versus C/I curves for all the CS/MCS has been
measured again for low and high Diversity scenarios. In case of significant changes in
the Link Adaptation threshold Tables the new values have been applied. In this way a
better throughput per user and in general an overall data capacity increase of the
network is reached.

10.5.1 Link Adaptation for GPRS


In the following, link adaptation features for GPRS service are described.
With GPRS the receiver operates only in type I ARQ mode (see "9.9.1.1 GPRS
Acknowledged Mode"); therefore link adaptation procedure is based only on this mode
of operation.

10.5.1.1 GPRS: Switching Points


As previously described, the exact values of the switching points depend on the real
network situation and are subjects of simulation and/or measurements.
For a first idea about the values to be expected, it is possible to interpret the following
simulations based on TU3 with ideal frequency hopping (see Fig. 10.4).

A30808-X3247-L24-5-7618 267
GPRS/EGPRS Global Description Information
System

Fig. 10.4 Gross Throughput Depending on CS and C/I (dB)

It is possible to estimate the 'ideal' switching points as follows:

CS1 <---> CS2 C/I=6.2 dB


CS2 <---> CS3 C/I=9.6 dB
CS3 <---> CS4 C/I=16.5 dB

However, switching points between coding schemes are defined in terms of BLER
thresholds. The corresponding BLER values are shown in Fig. 10.5.

268 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Fig. 10.5 BLER as Function of C/I (dB) for all GPRS Coding Schemes

For GPRS, the following BLER thresholds, e.g., can be defined:

BLER_CS1(CS1 ---> CS2) = 17% BLER_CS2(CS2 ---> CS1) = 43% C/I=6.2 dB


BLER_CS2(CS2 ---> CS3) = 14% BLER_CS3(CS3 ---> CS2) = 26% C/I=9.6 dB
BLER_CS3(CS3 ---> CS4) = 2% BLER_CS4(CS4 ---> CS3) = 17% C/I=16.5 dB

For example, if BLER goes below 17% while using CS1, then a change to CS2 will be
decided; if BLER goes over 43% while using CS2, then a change to CS1 will be decided.
A crosscheck e.g., for CS1<->CS2 provides approximately the same gross throughputs:

(1-0.17)*9.05 kbit/s=7.51 kbit/s (1-0.43)*13.4 kbit/s=7.64 kbit/s

It is also possible to see – depending on the wished QoS – that the hysteresis should
be more towards the more stable CS: in the example above, both CS have nearly the
same gross throughput, but with CS2, 43% of all blocks mus bet repeated at least once,
so the delay will be much higher than if one uses CS1. Therefore the '-' point should be
as close as possible to the '=' point.
When considering the net throughput, the maximum data rate values would become 8,
12, 14.4 and 20 kbits/s. The curves above (see Fig. 10.4) are re-scaled, each one by a
proper factor, and the 'ideal' switching points should be recalculated accordingly. These
switching points are reported in Tab. 10.3, and are the values that are contained in the
GPRS internal switching matrix. It is important to underline that for GPRS, only one
switching matrix exists, independently of the value given to the RAENV parameter.

CS1 CS2 CS3 CS4

CS1 <10%

Tab. 10.3 GPRS Thresholds with RAENV set to LOWDIV/HIGHDIV

A30808-X3247-L24-5-7618 269
GPRS/EGPRS Global Description Information
System

CS1 CS2 CS3 CS4

CS2 >50% <10%


CS3 >50% <1%
CS4 >30%

Tab. 10.3 GPRS Thresholds with RAENV set to LOWDIV/HIGHDIV

In Tab. 10.3, coding schemes written in the vertical direction represent the starting
coding schemes, whereas those written in the horizontal direction represent the arrival
ones. For example, watching at Tab. 10.3, to go from CS1 to CS2, the BLER value must
be less than 10%; to go from CS2 to CS1, the BLER value must be greater than 50%.
Let BLER(CSi-->CSi+1) be the (upgrade) switching point from CSi to CSi+1 and
BLER(CSi<--CSi+1) the corresponding (downgrade) switching point and
BLER(CSi=CSi+1) the BLER of the current coding scheme where both corresponding
coding schemes have with same C/I the same throughput.
Then the following must always be valid:

1) BLER(CSi-->CSi+1)<BLER(CSi=CSi+1) i=1..3
2) BLER(CSi=CSi+1)<=BLER(CSi<--CSi+1) i=1..3

10.5.1.2 “Quality Traps” Disadvantage


It would be desirable to enforce in the downgrade situation, a re-sending of the NACK
blocks in the new, more robust coding scheme.
However, this is forbidden by Specifications and it may lead to the following situation
(“quality trap”): if conditions worsen (or a too “high” CS was selected in the beginning)
before all NACK blocks could be resent successfully, remaining blocks will never be
transferred at all (conditions are too bad for transfer with current CS, but CS may not be
changed because blocks must be resent with the old CS).
The situation described above is very pessimistic. In more common cases, interference
i and fading conditions are variable enough, from block to block, to allow an RLC block
correct reception within a reasonable number of retransmissions. If p is the probability
of retransmission, each RLC block will be correctly received after an average 1/(1-p)
number of retransmissions.

Therefore, when radio conditions are bad and the link adaptation leads to switching to
a lower coding scheme, in progress retransmission will be in any case performed using
the ‘old’ coding scheme. As usual, when the number of retransmissions of a block
exceeds the N3101 value, the TBF is closed. If the TBF is re-opened within a time
configured by the STGTTLLIINF parameter, it will be re-opened with the last
commanded/used coding scheme, overtaking quality traps disadvantages (see
"10.5.3 Selection of the Candidate Initial Coding Scheme" to get more details about this
procedure).
For example, for transmission in the downlink direction, if the current coding scheme is
CS4 and link adaptation leads to switch to CS3 than:
– CS3 coding scheme is used to transmit new blocks
– CS4 (‘old’ coding scheme) is used for in progress retransmissions

270 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

When the number of retransmissions of a block exceeds the N3101 value, the TBF is
closed. If the TBF is re-opened within the STGTTLLIINF time, CS3 coding scheme (last
used coding scheme) will be used.
For example, for transmission in the uplink direction, if the current coding scheme is CS4
and link adaptation leads to switch to CS3 than:
– CS3 coding scheme is commanded (and used by the MS to transmit new blocks)
– CS4 (‘old’ coding scheme) is used by the MS for in progress retransmissions
When the number of retransmissions of a block exceeds the N3101 value, the TBF is
closed. If the MS again requires the TBF within the STGTTLLIINF time, CS3 coding
scheme will be used (last commanded coding scheme).

10.5.1.3 GPRS: Link Adaptation Algorithm


Speaking about GPRS service, link adaptation is based on the following features:
1. All of the signalling is done with CS1;
2. Data transfer is started with the highest available coding scheme; the following
considerations can be drawn:
– if the transmission to be started is a retry due to a previous “quality trap”
(see 10.5.1.2), the CS to be selected must be more stable than the one used
before. The PCU must store abnormal releases for a certain time, to recognize a
retry and adapt the CS for that case.
– as described in 10.5.1.2, the PCU holds in memory the value of the last used
coding scheme for a defined time; this coding scheme is the coding scheme to be
used when a new TBF is opened. If the time is elapsed (or if a cell reselection is
executed) the used coding scheme is those provided by the INICSCH parameter
(see "10.5.3 Selection of the Candidate Initial Coding Scheme").
3. Basing on received blocks, the BSC can evaluate the BLER value;
4. If conditions worsen, a downgrade to the lower CS is performed, but all of NACK
blocks must be successfully sent using the old coding scheme (i.e., after a change
from coding scheme A to coding scheme B, all of the not acknowledged blocks must
be resent using the coding scheme A); this may lead to the 'quality trap'. This
problem is solved only by expiration of a timer and reestablishing the transmission
with a more appropriate CS.
5. If conditions improve, an upgrade to the higher CS is performed, but all of NACK
blocks must successfully be sent using the old coding scheme. In uplink case, the
BSC informs the MS to change the coding scheme; in downlink direction, the BSC
starts sending blocks with the new coding scheme. The upgrade procedure depends
not only on the propagation conditions (i.e., C/I or BLER) but also on the available
resources. A switch (e.g., CS2 to CS3) is not allowed, if there is e.g., insufficient
transmission capacity on the Abis (e.g., a 32 kbit/s channel available for a CS3, see
"6.3 PCU Frames and Dynamic Allocation on the Abis Interface").

10.5.2 Link Adaptation for EGPRS


For EGPRS, up to nine modulation and coding schemes are defined; each scheme
belong to a certain family.
In general, according to the link quality, an initial Modulation and Coding Scheme (MCS)
is selected for an RLC block. For retransmissions, the same or another MCS from the
same family of MCSs can be selected. For example, if MCS7 is selected for the first

A30808-X3247-L24-5-7618 271
GPRS/EGPRS Global Description Information
System

transmission of an RLC block, any MCSs of the family B can be used for the retransmis-
sion.
In the type I ARQ mode (see "9.9.1.2 EGPRS Acknowledged Mode"), decoding of an
RLC Data Block is based solely on the prevailing transmission (i.e., erroneous blocks
are not stored). In the type II ARQ case, erroneous blocks are stored by the receiver and
a joint decoding with new transmissions is done. Link Adaptation procedure allows the
receiver to operate either in type I or type II hybrid ARQ mode.

10.5.2.1 EGPRS: Switching Points


With EGPRS, different MCSs allow different performance (throughput) as a function of
the C/I ratio (and the actual RF channel); Fig. 10.6 shows curves for TU3 with no
frequency hopping, for family A + MCS1 (without Incremental Redundancy).

Fig. 10.6 Simulation Results for Family A (+MCS1) without IR

It is possible to estimate the 'ideal' C/I switching points where the MCS should be
changed in order to maximize the net throughput. Referring to the previous case, and
assuming only using MCSs belonging to family A (plus MCS1), the 'ideal' switching
points could be as follows:

MCS1 <---> MCS3 C/I=1.5 dB


MCS3 <---> MCS6 C/I=7.5 dB
MCS6 <---> MCS9 C/I=16 dB

272 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

As described for GPRS, also with EGPRS the thresholds are given in terms of BLER,
since C/I values are difficult to estimate; some examples of defined BLER thresholds
could be as follows:

BLER(MCS1 ---> MCS3) = 61% BLER(MCS3 ---> MCS1) = 77% C/I=1.5 dB


BLER(MCS3 ---> MCS6) = 38% BLER(MCS6 ---> MCS3) = 69% C/I=7.5 dB
BLER(MCS6 ---> MCS9) = 24% BLER(MCS9 ---> MCS6) = 62% C/I=16 dB

The throughput is maximized changing the coding schemes according to these


threshold values. For example, if BLER goes below 38%, while using MCS3, then a
change to MCS6 will be decided; if BLER goes over 69%, while using MCS6, then a
change to MCS3 will be decided.
In real networks, 'ideal' switching points will depend on the actual RF scenario and it is
impossible to calculate such optimal values for each particular scenario.
Upgrade switching points (BLER(MCSx--->MCSy), 1<=x<y<=9) and downgrade
switching points (BLER(MCSy--->MCSx), 1<=x<y<=9) are adaptable to the radio envi-
ronment.
More precisely, these values are stored in pre-calculated matrix tables, one for each
possible RF environment; therefore:
– if RAENV is set to HIGHDIV, some threshold values are used
– if RAENV is set to LOWDIV, some other threshold values are used
With O&M, it is possible (by setting the RAENV parameter) to select the suitable matrix
table, containing all of the ideal switching points (downgrade/upgrade switching points
from/to all MCSs) for the particular RF scenario.
When IR is taken into account, different BLER threshold values should be considered:
BLER(MCSx_wIR--->MCSy_wIR), 1<=x<y<=9 as upgrade switching points and
BLER(MCSy_wIR--->MCSx_wIR), 1<=x<y<=9, as downgrade switching points. These
values are stored in matrix tables, one for each possible RF scenario.
Among all possible EGPRS coding schemes, the operator must define which sets of
coding schemes must be used in the cell, both in uplink and downlink direction (see
"10.5.2.2 EGPRS: Link Adaptation Algorithm" to know how to enable EGPRS coding
schemes in a cell). When configuring these sets, the following constraints are automat-
ically considered by the system:
a) MCS1 is always included in the selection: this is needed for signalling and due to the
fact that MCS1 is the only MCS that requires only one Abis timeslot;
b) If MCSx is implemented, all of the lower order MCSy (y<x) of the same family must
be implemented. This is needed because retransmissions may have to be
performed with a (lower order) MCS of the same family.
Link Adaptation is not restricted within a family. Only retransmissions must
i be performed using a coding scheme in the same family of the original one
(the one that was used for the first transmission of the radio block).

c) If more than one family is configured, considering a given MCSx in use, the general
rule to decide the upgrading/downgrading MCS is the following: the “upgrading”
MCS is the one characterized by the highest switching threshold (among the config-

A30808-X3247-L24-5-7618 273
GPRS/EGPRS Global Description Information
System

ured ones), while the “downgrading” MCS is the one characterized by the lowest
switching threshold (among the configured ones).
It may happen that the upgrading threshold is higher that the downgrading
i threshold. In this case one of the two conditions is always satisfied (implic-
itly this means that the current MCS is a “transition one” and the best
choice is to immediately switch to a new one). In case both conditions are
satisfied, the best choice is to switch to the upgrading MCS.

According to the chosen sets of coding schemes (in uplink or downlink direction),
different thresholds must be considered, since different coding schemes are selected.
Tab. 10.4 shows which thresholds are considered if, for instance, the user has enabled
FamilyA plus MCS1. Instead Tab. 10.5 shows which thresholds are considered if, for
instance, the user has enabled FamilyB plus MCS1.
In the tables, coding schemes written in vertical direction represent the starting coding
schemes, whereas those written in horizontal represent the arrival ones. Therefore, for
example, watching at Tab. 10.4, to go from MCS1 to MCS3 the BLER must be less than
a XX% value; to go from MCS3 to MCS1 the BLER must be greater than another XX%
value.
If more than one family is enabled, the possible switching points are those given by the
i sum of tables related to the single families.

Fam C Fam B Fam A + Fam C Fam B Fam A + Fam B Fam A Fam A


Fam A Fam A Padding
Padding Padding
MCS1 MCS2 MCS3 MCS4 MCS5 MCS6 MCS7 MCS8 MCS9
Fam C MCS1 <XX%
Fam B MCS2
Fam A + MCS3 >XX% <XX%
Fam A
Padding
Fam C MCS4
Fam B MCS5
Fam A + MCS6 >XX% <XX%
Fam A
Padding
Fam B MCS7
Fam A MCS8
Padding
Fam A MCS9 >XX%

Tab. 10.4 Thresholds to be used if Family A plus MCS1 are chosen

274 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Fam C Fam B Fam A + Fam C Fam B Fam A + Fam B Fam A Fam A


Fam A Fam A Padding
Padding Padding
MCS1 MCS2 MCS3 MCS4 MCS5 MCS6 MCS7 MCS8 MCS9
Fam C MCS1 <XX%
Fam B MCS2 >XX% <XX%
Fam A + MCS3
Fam A
Padding
Fam C MCS4
Fam B MCS5 >XX% <XX%
Fam A + MCS6
Fam A
Padding
Fam B MCS7 >XX%
Fam A MCS8
Padding
Fam A MCS9

Tab. 10.5 Thresholds to be used if Family B plus MCS1 are chosen

If all EGPRS coding schemes are enabled, according to radio environment (i.e., RAENV
parameter setting) and IR, four different threshold settings are foreseen:
a) RAENV set to LOWDIV and IR is working (see Tab. 10.6)
b) RAENV set to LOWDIV and IR is not working (see Tab. 10.7)
c) RAENV set to HIGHDIV and IR is working (see Tab. 10.8)
d) RAENV set to HIGHDIV and IR is not working (see Tab. 10.9)

To MCS1 MCS2 MCS3 MCS4 MCS5 MCS6 MCS7 MCS8 MCS9


From
MCS1 <35% <35% <30%
MCS2 >60% <30% <25% <35%
MCS3 >70% >60% <20% <35% <30%
MCS4 >70% >60% >45% <40% <40%
MCS5 >70% >65% >70% <40% <45%
MCS6 >70% >75% >70% <40% <40% <40%
MCS7 >80% >70% <30% <30%
MCS8 >75% >75% <15%
MCS9 >80% >80% >80%

Tab. 10.6 EDGE with Incremental Redundancy working and RAENV set to "LOWDIV"

A30808-X3247-L24-5-7618 275
GPRS/EGPRS Global Description Information
System

To MCS1 MCS2 MCS3 MCS4 MCS5 MCS6 MCS7 MCS8 MCS9


From
MCS1 <35% <35% <30%
MCS2 >60% <30% <25% <35%
MCS3 >70% >60% <20% <35% <30%
MCS4 >70% >50% >45% <40% <40%
MCS5 >70% >65% >70% <40% <25%
MCS6 >70% >75% >70% <30% <25% <15%
MCS7 >65% >60% <30% <15%
MCS8 >65% >50% <15%
MCS9 >65% >45% >30%

Tab. 10.7 EDGE with Incremental Redundancy not working and RAENV set to "LOWDIV"

To MCS1 MCS2 MCS3 MCS4 MCS5 MCS6 MCS7 MCS8 MCS9


From
MCS1 <15% <10% <10%
MCS2 >45% <3% <1% <20%
MCS3 >50% >35% <5% <40% <35%
MCS4 >55% >40% >25% <45% <45%
MCS5 >70% >70% >70% <15% <30%
MCS6 >70% >70% >45% <5% <15% <40%
MCS7 >70% >45% <5% <5%
MCS8 >60% >25% <15%
MCS9 >75% >35% >50%

Tab. 10.8 EDGE with Incremental Redundancy working and RAENV set to "HIGHDIV"

To MCS1 MCS2 MCS3 MCS4 MCS5 MCS6 MCS7 MCS8 MCS9


From
MCS1 <15% <3% <1%
MCS2 >45% <3% <1% <10%
MCS3 >50% >35% <5% <50% <35%
MCS4 >55% >40% >25% <90% <75%
MCS5 >60% >80% >90% <15% <2%
MCS6 >70% >90% >45% <5% <2% <2%
MCS7 >55% >45% <5% <5%

Tab. 10.9 EDGE with Incremental Redundancy not working and RAENV set to "HIGHDIV"

276 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

MCS8 >50% >25% <15%


MCS9 >55% >35% >50%

Tab. 10.9 EDGE with Incremental Redundancy not working and RAENV set to "HIGHDIV"

To summarize, the basic ideas are as follows:


• link Adaptation is based on BLER measurements
• different BLER thresholds are needed to take into account type I ARQ and type II
Hybrid ARQ cases
• only a subset of MCSs should be used; especially due to the fact that is quite difficult
to define consistent thresholds among all of the possible MCSs.

10.5.2.2 EGPRS: Link Adaptation Algorithm


Regarding the EGPRS link adaptation algorithm, some differences exist between uplink
and downlink directions; these differences arise from the following situations:
a) there are EGPRS mobile stations that are unable to use PSK modulation in uplink
direction, but only the GMSK one (see "9.1 Mobile Stations for Packet Switched
Services");
b) incremental redundancy is not supported in uplink direction.

Uplink Direction
The operator can configure which sets of coding schemes can be used in uplink direc-
tion. At least two sets of available MCSs must be enabled, one for 8PSK transmit
capable mobiles and the other one for GMSK-only transmit capable mobiles.
To enable sets of coding schemes, a parameter is given for each family. The parameters
are as follows (remember that in uplink direction, the Incremental Redundancy is not
implemented):
• EMFA1UNIR8PSK (enMcsFamAMcs1UplinkWoutIncrRed8Psk): enables MCS
belonging to FamilyA and MCS1 to be used, if MS support 8PSK modulation in the
Uplink;
• EMFAP1UNIR8PSK
(enableMcsFamilyApMcs1UplinkWithoutIncrementalRedundancy8Psk): enables
MCS belonging to FamilyA padding and MCS1 to be used, if MS support 8PSK
modulation in the Uplink case;
• EMFB1UNIR8PSK (enMcsFamBMcs1UplinkWoutIncRed8Psk): enables MCS
belonging to Family B and MCS1 to be used, if MS support 8PSK modulation in the
Uplink EGPRS TBF;
• EMFCUNIR8PSK (enMcsFamCUplinkWoutIncRed8Psk): enables MCS belonging
to Family C and MCS1 to be used, if MS support 8PSK modulation in the Uplink
case;
• EMFGUNIR8PSK (enMcsFamGmskUplinkWoutIncRed8Psk): enables MCS
belonging to Family Gmsk to be used, if MS supports 8PSK modulation in the Uplink
case;
• EMFCUNIRGMSK (enMcsFamCUplinkWoutIncrRedGmsk): enables MCS
belonging to Family C to be used, if MS does not support 8PSK modulation in the
Uplink case;

A30808-X3247-L24-5-7618 277
GPRS/EGPRS Global Description Information
System

• EMFGUNIRGMSK (enMcsFamGmskUplinkWoutIncrRedGmsk): enables MCS


belonging to Family Gmsk to be used, if MS does not support 8PSK modulation in
the Uplink case.
Family GMSK is constituted by MSCs that can be used from a mobile station supporting,
i in uplink direction, the GMSK modulation only. The coding schemes belonging to Family
GMSK are: MCS1, MCS2, MCS3, and MCS4; these coding schemes use the GMSK
modulation (see "3.1 GPRS and EGPRS Modulation Principles").

The operator can also define the initial MCS to be used as default in uplink direction; if
no information is available about a MS in a cell, the defined MCSs will be used (see
"10.5.3 Selection of the Candidate Initial Coding Scheme").
The IMCSULNIR8PSK attribute suggests the MCS to be used in uplink direction if the
MS supports the 8 PSK modulation in this direction; the IMCSULNIRGMSK attribute
suggests the MCS to be used in uplink direction if the MS supports only the GMSK
modulation in this direction.
The link adaptation algorithm in uplink direction works as follows:
1. The initial Modulation and Coding Scheme is decided. In the absence of Abis
congestion, the initial MCS will be IMCSULNIR8PSK (or IMCSULNIRGMSK for
GMSK mobiles), unless some information is available about the last MCS used for
a previous UL TBF characterized by the same TLLI. In this case, the initial MCS of
the new TBF will be set equal to the last MCS of the previous one (see 10.5.3);
2. Once the connection is established, BLER is continuously updated at the PCU (each
20 ms) by checking whether or not RLC blocks have been carefully received; the
filtering period can be defined, in number of radio blocks, by the BLERAVEUL
parameter;
3. Once the initial filtering period has elapsed (e.g., after 100 radio blocks if BLER-
AVEUL is set to UNIT100), BLER is continuously monitored. Each time (i.e., for each
received block) BLER is checked and tested against the appropriate thresholds; if
MCSx is the actual MCS, MCSy the next available one and MCSz the previous avail-
able one, the appropriate thresholds are:

Up_th= BLER(MCSx--->MCSy) upgrade threshold


Dn_th= BLER(MCSx--->MCSz) downgrade threshold

4. If actual BLER falls below the upgrade threshold (Up_th), the algorithm switches to
the next (less protected) available MCS; if it exceeds the downgrade threshold
(Dn_th), the algorithm switches to the previous (more protected) available MCS.
When upgrading to a less protected MCS, Abis availability should be checked, see
i "6.3 PCU Frames and Dynamic Allocation on the Abis Interface" and "5.3.4.2 Upgrade
of Abis Resources".

Downlink direction
In the downlink, direction incremental redundancy is assumed to always be enabled,
since it is mandatory for EGPRS mobiles.
The operator can configure which sets of coding schemes can be used in the downlink
direction. EGPRS mobiles will be able to receive 8PSK modulated signals, therefore at
least one family of available MCSs must be enabled (all of the MSs are 8PSK receive
capable).

278 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

To enable sets of coding schemes, a parameter for each family is given. The following
are the parameters (remember that in downlink direction, the Incremental Redundancy
is always supported by MSs):
• EMCSFAMA1DL (enMcsFamAMcs1Downlink): enables MCS belonging to Family A
and MCS1 to be used, in Downlink case;
• EMCSFAMAP1DL (enMcsFamAPaddingMcs1Downlink): enables MCS belonging
to Family A padding and MCS1 to be used, in Downlink case;
• EMCSFAMB1DL enMcsFamBMcs1Downlink): enables MCS belonging to Family B
and MCS1 to be used, in Downlink case;
• EMCSFAMCDL (enMcsFamCDownlink): enables MCS belonging to Family C to be
used, in Downlink case;
• EMCSFAMGDL (enableMcsFamilyGmskDownlink): enables MCS belonging to
Family GSMK to be used, in Downlink case.
The operator can also define, by the INIMCSDL attribute, the initial MCS to be used as
the default in the downlink direction; if no information about the MS in a cell is available,
the suggested MCSs will be used (see "10.5.3 Selection of the Candidate Initial Coding
Scheme").
The link adaptation algorithm in the downlink direction works as follows:
1. The initial Modulation and Coding Scheme is decided. In the absence of Abis
congestion, the initial MCS will be INIMCSDL, unless some information is available
about the last MCS used for a previous DL TBF characterized by the same TLLI. In
this case, the initial MCS of the new TBF will be set equal to the last MCS of the
previous one (see 10.5.3);
2. Once the connection is established, BLER is updated at the PCU with the informa-
tion provided by the EGPRS PACKET DOWNLINK ACK/NACK MESSAGE, reported
by the MS upon periodic request from the network (let k be the reporting instant); the
filtering period can be defined, in number of radio blocks, by the BLERAVEDL
parameter;
3. When an EGPRS PACKET DOWNLINK ACK/NACK message is received (i.e., at the
instant k), the MS OUT OF MEMORY bit is checked to verify if no more memory for
incremental redundancy procedure is available at the MS. From the MS OUT OF
MEMORY bit, the IR_status_k variable is derived, providing information about the
efficiency of incremental redundancy at the MS at a specific instant k:
– IR is considered as "not-properly working” when IR_status_k<0.5
– IR is considered as "properly working” when IR_status_k>0.5
4. BLER is continuously monitored; each time an EGPRS PACKET DOWNLINK
ACK/NACK is received, BLER is checked and tested against the appropriate thresh-
olds; if MCSx is the actual MCS, MCSy the next available one and MCSz the
previous available one, the appropriate thresholds are:
– if IR was perfect (no memory size limitations, etc.), the appropriate thresholds
would be:

Up_th_k= BLER(MCSx_wIR--->MCSy_wIR) upgrade threshold


Dn_th_k= BLER(MCSx_wIR--->MCSz_wIR) downgrade threshold

– if IR did not work at all, the appropriate thresholds would be:

A30808-X3247-L24-5-7618 279
GPRS/EGPRS Global Description Information
System

Up_th_k= BLER(MCSx--->MCSy) upgrade threshold


Dn_th_k= BLER(MCSx--->MCSz) downgrade threshold

– generally (IR efficiency given by IR_statusk) the appropriate thresholds would be


something between thresholds defined above:

Up_th_k= (1-IR_status_k)*BLER(MCSx--->MCSy)+ upgrade


IR_status_k*BLER(MCSx_wIR--->MCSy_wIR) threshold
Dn_th_k= (1-IR_status_k)*BLER(MCSx--->MCSz)+ downgrade
IR_status_k*BLER(MCSx_wIR--->MCSz_wIR) threshold

5. If actual BLER falls below the upgrade threshold (Up_th_k), the algorithm switches
to the next (less protected) available MCS; if it exceeds the downgrade threshold
(Dn_th_k), the algorithm switches to the previous (more protected) available MCS.
When upgrading to a less protected MCS, Abis availability should be checked, see
i "6.3 PCU Frames and Dynamic Allocation on the Abis Interface" and "5.3.4.2 Upgrade
of Abis Resources".

10.5.3 Selection of the Candidate Initial Coding Scheme


The Initial Coding Scheme is the coding scheme to be applied, when a new TBF starts.
As previously described, the user can configure the coding scheme to be used when a
new data transmission starts both for GPRS and EGPRS services:
• for GPRS service, the INICSCH parameter is used;
• for EGPRS service, in uplink direction, as described in "9.1 Mobile Stations for
Packet Switched Services", not all the of EDGE mobile stations support the 8PSK
modulation, therefore:
– the IMCSULNIR8PSK attribute suggests that the MCS to be used in uplink direc-
tion if the MS supports the 8 PSK modulation in this direction;
– the IMCSULNIRGMSK attribute suggests that the MCS to be used in uplink direc-
tion if the MS supports only the GMSK modulation in this direction;
• for EGPRS service, in downlink direction, all of the EDGE MS are obliged to support
the 8 PSK modulation, so the INIMCSDL attribute suggests the MCS to be used in
downlink direction for all the EGPRS mobiles.
These values are used to choose the initial coding scheme, only when the PCU does
not have valid information of the coding scheme to use when the TBF starts.
In fact, the PCU holds into memory, for each mobile station, the last coding scheme (CS
or MCS) used in Uplink or Downlink direction for TBFs associated to the MS. These
values are refreshed at the end of each TBF and cleared from memory if either a timer
defined by the STGTTLLIINF (storageTimeTLLIInfo) parameter expires or if a cell rese-
lection is performed. These values of coding schemes are called “historical” coding
schemes; configured by O&M coding schemes will be used only if no historical values
are available at the PCU side.
Besides the coding scheme, the PCU also holds in memory, for the same time, the
i “historical BLER”, i.e., the last measured BLER. This value is used to assign radio
resources to the new TBF (see "5.3.3 Management of Incoming GPRS/EGPRS
Requests").

280 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

When a new TFB starts, the Candidate Initial Coding Scheme must be selected:
a) for GPRS capable mobiles, only the ‘candidate initial CS’ must be calculated
b) for EGPRS capable mobiles, both the ‘candidate initial MCS’ and the ‘candidate
initial CS’ must be calculated. In fact, only after the Resource Allocation procedure
(see "5.3.3.1 PCU Algorithm" and "5.3.3.2 TDPC Algorithm") will be clear which TBF
mode (GPRS or EGPRS) is to be used.
Therefore, the candidate initial coding scheme is selected in the order looking at the
following.
For GPRS capable mobiles:
– historical CS, if available
– configured initial CS, if historical CS is not available
For EGPRS capable mobiles:
a) canditate initial MCS:
– based on historical MCS, if any
– based on historical CS, if available (see Tab. 10.11)
– configured initial MCS, if historical CS/MCS are not available
b) canditate initial CS:
– historical CS, if available
– based on historical MCS, if available (see Tab. 10.10)
– configured initial CS, if historical CS/MCS are not available
Besides, for EGPRS service it is important to remember that the operator can configure
separately:
– MCS families to be used in downlink transmission
– MCS families to be used in uplink transmission for 8PSK capable mobiles
– MCS families to be used in uplink transmission for GMSK capable mobiles
As a consequence, it could happen that an available (historical) MCS cannot be directly
usable for the new TBF to be set up, because e.g., the user has changed the value of
O&M parameters. In this case, the rule to select the candidate initial MCS is: take the
highest configured MCS less or equal to the available historical MCS.
The following tables show the rules to decide the candidate initial coding scheme for the
following:
– a GPRS TBF mode, when the last coding scheme was stored for an EGPRS TBF
mode (see Tab. 10.10).
– a EGPRS TBF mode, when the last coding scheme was stored for a GPRS TBF
mode (see Tab. 10.11);

Historical MCS Candidate CS

MCS1 CS1
MCS2 CS2
MCS3 CS3
MCS4 or higher MCSs CS4

Tab. 10.10 Candidate Initial Coding Scheme for a GPRS TBF when the Historical
Coding Scheme is related to an EGPRS TBF

A30808-X3247-L24-5-7618 281
GPRS/EGPRS Global Description Information
System

Historical CS Candidate MCS

CS1 -MCS1
CS2 -MCS2 if FamilyB is configured
-MCS1 otherwise
CS3 -MCS3 if FamilyA is configured
-MCS2 if FamilyB is configured
-MCS1 otherwise
CS4 DL or UL TBF for fully 8PSK UL TBF for GMSK capable
capable mobiles mobiles
-Configured Initial MCS if it is -MCS4 if FamilyC is configured
upper than MCS4 -MCS3 if FamilyA is configured
-MCS4 if FamilyC is configured -MCS2 if FamilyB is configured
-MCS3 if FamilyA is configured -MCS1 otherwise
-MCS2 if FamilyB is configured
-MCS1 otherwise

Tab. 10.11 Candidate Initial Coding Scheme for an EGPRS TBF when the Historical
Coding Scheme is related to a GPRS TBF

282 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

11 Database Parameters and Objects


The current chapter contains four tables:
– In the first table the user finds, in the alphabetical order, all of the parameters, related
only to GPRS/EGPRS, that are discussed in the manual. For each parameter the
user finds one or more links to the chapters of the manual where the parameter is
described and also a link to the title of Feature Sheets (or Change Requests) that
introduce or describe the parameter in Siemens technology;
– In the second one the user finds, in the alphabetical order, the unspecific
GPRS/EGPRS parametersthat are discussed in the manual since they are also
related to GPRS/EGPRS . For each parameter, he/she finds one or more links to the
chapters of the manual where the parameter is described; besides, starting from
parameters of BR5.5 release onwards, a link to the features that describe the param-
eter is executed.
– In the third one the user finds, in the alphabetical order, all of the database objects
that are related only to GPRS/EGPRS. For each object he/she finds the link to
Feature Sheets (or Change Requests) that introduce the object in Siemens tech-
nology;
– In the fourth one, the user finds, in the alphabetical order, all of the unspecific
GPRS/EGPRS database objects that are also involved in GPRS/EGPRS. For each
object he finds the link to the chapters of the manual that describe it; besides,
starting from parameters of BR5.5 release onwards, a link to the features that
describe the parameter is executed.

Parameter Feature/CR Chapters

ABUTYP FSH 0720 "9.8.2.1 8 Bit or 11 Bit Uplink Access"


"9.8.2.4 TBF Establishment for EDGE Mobile
Stations"
ACCEPTGDEGR FSH 0516 "5.3.4.1 Upgrade of Radio Resources"
ALPHA FSH 0720 "10.4.1 Power Control Algorithm"
"10.4.2.2 Packet Transfer Mode: Measure-
ments for Power Control"
BAF FSH 0720 "7.1 Physical Layer"
BEPAVGP FSH 0420 NOT USED IN BR7.0
BER FSH 0720 "7.1 Physical Layer"
BLERAVEDL FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
BLERAVEUL FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
BNDWIDTDD FSH 0418 "10.2.3 Handling of UMTS Neighboring Cells"
BPAGCHR FSH 0720 "4.5.2 PDCH Carrying both PBCCH and
PCCCH"
"9.8.3.2 Discontinuous Reception"
"4.5.3 PDCH Carrying PCCCH"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

A30808-X3247-L24-5-7618 283
GPRS/EGPRS Global Description Information
System

Parameter Feature/CR Chapters

BPRACHR FSH 0720 "4.5.2 PDCH Carrying both PBCCH and


PCCCH"
"4.5.3 PDCH Carrying PCCCH"
BSCDVMA FSH 0720 "9.9.3.4 Release of an Uplink TBF"
BSPBBLK FSH 0720 "4.5.2 PDCH Carrying both PBCCH and
PCCCH"
"9.8.3.2 Discontinuous Reception"
BVCBHIPER FSH 0514 "7.3.3 SGSN-BSS Flow Control"
BVCBMAPER FSH 0514 "7.3.3 SGSN-BSS Flow Control"
BVCBLPER FSH 0514 "7.3.3 SGSN-BSS Flow Control"
BVCBSPPER FSH 0514 "7.3.3 SGSN-BSS Flow Control"
C31H FSH 0720 "10.1.3 Cell Re-selection Algorithm"
C32QUAL FSH 0720 "10.1.3 Cell Re-selection Algorithm"
CACKTYP FSH 0720 "9.8.5 Polling Procedures"
CR - X617
CODE FSH 0720 "7.1 Physical Layer"
CRC FSH 0720 "7.1 Physical Layer"
CRESELTRHSOUT FSH 0418 "10.3.2 GPRS/EGPRS Traffic Control
Strategy"
CRESELTRSHINP FSH 0418 "10.3.2 GPRS/EGPRS Traffic Control
Strategy"
CSCH3CSCH4SUP FSH 0419 "4.2.1 GPRS Channel Coding"
DRXTMA FSH 0720 "9.8.3.2 Discontinuous Reception"
CR - F190
CR - X617
EBCCHTRX FSH 0420 "5.1.3 Aspects Related to Carrier Configura-
tion"
"5.3.3 Management of Incoming
GPRS/EGPRS Requests"
EEDGE FSH 0420 "5.1 Enabling Packet Switched Services in a
CR - X0158 Cell"

EGPLGPONETS FSH 0420 "9.9.3.1 Uplink TBF Using the Acknowledged


Mode"
"9.9.4.1 Acknowledged and Unacknowledged
Modes on Downlink TBFs"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

284 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Parameter Feature/CR Chapters

EGPLGPTWOTS FSH 0420 "9.9.3.1 Uplink TBF Using the Acknowledged


Mode"
"9.9.4.1 Acknowledged and Unacknowledged
Modes on Downlink TBFs"
EGPLGPTHREETS FSH 0420 "9.9.3.1 Uplink TBF Using the Acknowledged
Mode"
"9.9.4.1 Acknowledged and Unacknowledged
Modes on Downlink TBFs"
EGPLGPFOURTS FSH 0420 "9.9.3.1 Uplink TBF Using the Acknowledged
Mode"
"9.9.4.1 Acknowledged and Unacknowledged
Modes on Downlink TBFs"
EGPLGPFIVETS FSH 0420 "9.9.3.1 Uplink TBF Using the Acknowledged
Mode"
"9.9.4.1 Acknowledged and Unacknowledged
Modes on Downlink TBFs"
EGPLGPSIXTS FSH 0420 "9.9.3.1 Uplink TBF Using the Acknowledged
Mode"
"9.9.4.1 Acknowledged and Unacknowledged
Modes on Downlink TBFs"
EGPLGPSEVENTS FSH 0420 "9.9.3.1 Uplink TBF Using the Acknowledged
Mode"
"9.9.4.1 Acknowledged and Unacknowledged
Modes on Downlink TBFs"
EGPLGPEIGHTTS FSH 0420 "9.9.3.1 Uplink TBF Using the Acknowledged
Mode"
"9.9.4.1 Acknowledged and Unacknowledged
Modes on Downlink TBFs"
EGPRS FSH 0420 "5.1 Enabling Packet Switched Services in a
CR - X0158 Cell"
EGWSONETS FSH 0420 "9.9.1.2 EGPRS Acknowledged Mode"
EGWSTWOTS FSH 0420 "9.9.1.2 EGPRS Acknowledged Mode"
EGWSTHREETS FSH 0420 "9.9.1.2 EGPRS Acknowledged Mode"
EGWSFOURTS FSH 0420 "9.9.1.2 EGPRS Acknowledged Mode"
EGWSFIVETS FSH 0420 "9.9.1.2 EGPRS Acknowledged Mode"
EGWSSIXTS FSH 0420 "9.9.1.2 EGPRS Acknowledged Mode"
EGWSSEVENTS FSH 0420 "9.9.1.2 EGPRS Acknowledged Mode"
EGWSEIGHTTS FSH 0420 "9.9.1.2 EGPRS Acknowledged Mode"
ELKADPT FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

A30808-X3247-L24-5-7618 285
GPRS/EGPRS Global Description Information
System

Parameter Feature/CR Chapters

EMCSFAMA1DL FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"


EMCSFAMAP1DL FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
EMCSFAMB1DL FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
EMCSFAMCDL FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
EMCSFAMGDL FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
EMFA1UNIR8PSK FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
EMFAP1UNIR8PSK FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
EMFB1UNIR8PSK FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
EMFCUNIR8PSK FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
EMFCUNIRGMSK FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
EMFGUNIR8PSK FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
EMFGUNIRGMSK FSH 0444 "10.5.2.2 EGPRS: Link Adaptation Algorithm"
ESUP FSH 0420 "5.1.2 Enabling EGPRS Service in the Cell"
FDDGQO CR - X482 "10.2.2 GSM-UMTS Re-selection Algorithm:
Packet Switched Case"
FRSTD FSH 0720 "7.1 Physical Layer"
GAMMA ---- "10.4.1 Power Control Algorithm"
"10.4.2.2 Packet Transfer Mode: Measure-
ments for Power Control"
GASTRABISTH FSH 0516 "5.3.2.4 Switching between VA and HA
according to Abis Interface Conditions"
"5.3.2.5 Allocation of Resources"
"5.3.3.2 TDPC Algorithm"
"5.3.4.2 Upgrade of Abis Resources"
GASTRTH FSH 0503 "5.3.2.3 Switching between VA and HA
According to Radio Conditions"
"5.3.2.5 Allocation of Resources"
"5.3.4.1 Upgrade of Radio Resources"
GDCH (ex GCCH) FSH 0720 "5.2 Configuration of GPRS Channels in a
FSH 0457 Cell"
FSH 0503
"5.3.2.3 Switching between VA and HA
CR - X706
According to Radio Conditions"
"9.8.3.2 Discontinuous Reception"
GCELLRESH FSH 0720 "10.1.3 Cell Re-selection Algorithm"
GFDDMURREP FSH 0418 "10.3.1.1 Measurement Reporting"
GFDDREPQTY FSH 0418 "10.3.1.1 Measurement Reporting"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

286 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Parameter Feature/CR Chapters

GHCSPC FSH 0720 "10.1.2.2 C31 Criterion"


CR - F205
"10.1.2.3 C32 Criterion"
"10.1.4 Management of GPRS/EGPRS
Neighboring Cells"
"10.1.4.3 Configuration of an Adjacent Cell
with GSUP= TRUE"
GHCSTH FSH 0720 "10.1.2.2 C31 Criterion"
CR - F205
"10.1.4 Management of GPRS/EGPRS
Neighboring Cells"
"10.1.4.3 Configuration of an Adjacent Cell
with GSUP= TRUE"
GLK FSH 0720 "7.1 Physical Layer"
"7.2.1.1 Examples of Addressing"
GMANMSAL FSH 0720 "4.3.3 Multiplexing MSs on the same PDCH:
Configuration"
"4.4.3 Packet Data Traffic Channel (PDTCH)"
"5.2 Configuration of GPRS Channels in a
Cell"
"5.3.2 Horizontal/Vertical Allocation Strate-
gies"
"5.3.2.1 Vertical Allocation Strategy (VA)"
"5.3.3.1 PCU Algorithm"
"10.3.2.1 Network Controlled Cell Reselection
Algorithm for Traffic Control Strategy"
GMANPAL FSH 0720 REMOVED IN BR6.0
CR - X232
GMANPRES CR - F187 "5.2 Configuration of GPRS Channels in a
(restored in BR7.0) Cell"
"5.3.2.1 Vertical Allocation Strategy (VA)"
"5.3.2.3 Switching between VA and HA
According to Radio Conditions"
"5.3.3.2 TDPC Algorithm"
"7.3.3 SGSN-BSS Flow Control"
GMAPERTCHRES CR - X706 REMOVED IN BR7.0

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

A30808-X3247-L24-5-7618 287
GPRS/EGPRS Global Description Information
System

Parameter Feature/CR Chapters

GMSTXPMAC FSH 0720 "10.1.2.1 GPRS/EGPRS Path Loss Criterion


FSH 0418 (C1 Criterion)"
"10.1.4 Management of GPRS/EGPRS
Neighboring Cells"
"10.1.4.3 Configuration of an Adjacent Cell
with GSUP= TRUE"
"10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
"10.3.2.1 Network Controlled Cell Reselection
Algorithm for Traffic Control Strategy"
GNMULBAC FSH 0418 "10.1.1 Measurements for Cell Selection and
Re-selection"
"10.3.1.1 Measurement Reporting"
GPATH FSH 0720 "9.8.2 TBF Establishment Initiated by the MS
on CCCH/PCCCH"
GPDPDTCHA FSH 0503 "5.2 Configuration of GPRS Channels in a
CR - X617 Cell"
"5.3.2.1 Vertical Allocation Strategy (VA)"
"5.3.2.3 Switching between VA and HA
According to Radio Conditions"
"7.3.3 SGSN-BSS Flow Control"
GPENTIME FSH 0720 "10.1.2.2 C31 Criterion"
CR - F205
"10.1.2.3 C32 Criterion"
"10.1.4 Management of GPRS/EGPRS
Neighboring Cells"
"10.1.4.3 Configuration of an Adjacent Cell
with GSUP= TRUE"
GRESOFF FSH 0720 "10.1.2.2 C31 Criterion"
CR - F205
"10.1.2.3 C32 Criterion"
CR - X617
"10.1.4 Management of GPRS/EGPRS
Neighboring Cells"
"10.1.4.3 Configuration of an Adjacent Cell
with GSUP= TRUE"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

288 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Parameter Feature/CR Chapters

GRXLAMI FSH 0720 "10.1.2.1 GPRS/EGPRS Path Loss Criterion


FSH 0418 (C1 Criterion)"
"10.1.4 Management of GPRS/EGPRS
Neighboring Cells"
"10.1.4.3 Configuration of an Adjacent Cell
with GSUP= TRUE"
"10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
"10.3.2.1 Network Controlled Cell Reselection
Algorithm for Traffic Control Strategy"
GS FSH 0720 "9.8.2.6 Uplink Access on PRACH (Access
Persistence Control)"
GSUP (TRX object) FSH 0512 "5.1.1 Enabling GPRS Service in the Cell"
"5.1.2 Enabling EGPRS Service in the Cell"
"5.2 Configuration of GPRS Channels in a
Cell"
GSUP (ADJC object) FSH 0720 "10.1.4 Management of GPRS/EGPRS
Neighboring Cells"
"10.1.4.3 Configuration of an Adjacent Cell
with GSUP= TRUE"
"10.1.4.4 Configuration of an Adjacent Cell
with GSUP= FALSE"
GTDDMURREP FSH 0418 "10.3.1.1 Measurement Reporting"
GTEMPOFF FSH 0720 "10.1.2.2 C31 Criterion"
CR - F205
"10.1.2.3 C32 Criterion"
"10.1.4 Management of GPRS/EGPRS
Neighboring Cells"
"10.1.4.3 Configuration of an Adjacent Cell
with GSUP= TRUE"
GTS FSH 0720 "7.1 Physical Layer"
"7.2.1.1 Examples of Addressing"
GTXINT FSH 0720 "9.8.2.6 Uplink Access on PRACH (Access
Persistence Control)"
GUMTSSRHPRI FSH 0418 "10.1.1 Measurements for Cell Selection and
Re-selection"
"10.2 Cell Re-selection from
GSM/GPRS/EGPRS Network to UMTS
Network"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

A30808-X3247-L24-5-7618 289
GPRS/EGPRS Global Description Information
System

Parameter Feature/CR Chapters

IMCSULNIR8PSK FSH 0444 "4.2.2 EGPRS Channel Coding"


"10.5.2.2 EGPRS: Link Adaptation Algorithm"
"10.5.3 Selection of the Candidate Initial
Coding Scheme"
IMCSULNIRGMSK FSH 0444 "4.2.2 EGPRS Channel Coding"
"10.5.2.2 EGPRS: Link Adaptation Algorithm"
"10.5.3 Selection of the Candidate Initial
Coding Scheme"
INIBLER FSH 0516 "5.3.3.1 PCU Algorithm"
"10.5.3 Selection of the Candidate Initial
Coding Scheme"
INICSCH FSH 0720 "4.2.1 GPRS Channel Coding"
"10.5.1.3 GPRS: Link Adaptation Algorithm"
"10.5.3 Selection of the Candidate Initial
Coding Scheme"
INIMCSDL FSH 0444 "4.2.2 EGPRS Channel Coding"
"10.5.2.2 EGPRS: Link Adaptation Algorithm"
"10.5.3 Selection of the Candidate Initial
Coding Scheme"
LOWBER FSH 0720 "7.1 Physical Layer"
MNTBMASK CR - X1869 "4.2.1 GPRS Channel Coding"
MSBHIPER FSH 0514 "7.3.3 SGSN-BSS Flow Control"
MSBMAPER FSH 0514 "7.3.3 SGSN-BSS Flow Control"
MSBLPER FSH 0514 "7.3.3 SGSN-BSS Flow Control"
MSBSPPER FSH 0514 "7.3.3 SGSN-BSS Flow Control"
N3101 FSH 0720 "9.9.3.3 Anomalies During an Uplink TBF"
CR - X617
FSH 0444 "10.5.1.2 “Quality Traps” Disadvantage"
N3103 FSH 0720 "9.9.3.4 Release of an Uplink TBF"
N3105 FSH 0720 "9.9.4.1 Acknowledged and Unacknowledged
Modes on Downlink TBFs"
N391 FSH 0720 "7.2.1.3 Procedures for PVCs"
N392 FSH 0720 "7.2.1.3 Procedures for PVCs"
N393 FSH 0720 "7.2.1.3 Procedures for PVCs"
NAVGI FSH 0720 "10.4.2.3 Derivation of Channel Quality
Reports"
NBVCBR FSH 0720 "7.3.1.1 BVC Procedures"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

290 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Parameter Feature/CR Chapters

NBVCUR FSH 0720 "7.3.1.1 BVC Procedures"


NBVCRR FSH 0720 "7.3.1.1 BVC Procedures"
NCC1TH FSH 0418 "10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
NCC1THADJC FSH 0418 "10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
"10.3.2.1 Network Controlled Cell Reselection
Algorithm for Traffic Control Strategy"
NCGPENTIME FSH 0418 "10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
"10.3.2.1 Network Controlled Cell Reselection
Algorithm for Traffic Control Strategy"
NCGRESOFF FSH 0418 "10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
"10.3.2.1 Network Controlled Cell Reselection
Algorithm for Traffic Control Strategy"
NCGPENTIME FSH 0418 "10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
"10.3.2.1 Network Controlled Cell Reselection
Algorithm for Traffic Control Strategy"
NCRARESH FSH 0418 "10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
"10.3.2.1 Network Controlled Cell Reselection
Algorithm for Traffic Control Strategy"
NCRESELFLAG FSH 0418 "10.3.1 Network Controlled Cell Reselection"
"10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
"10.3.2 GPRS/EGPRS Traffic Control
Strategy"
NCSARA FSH 0418 "10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
"10.3.2.1 Network Controlled Cell Reselection
Algorithm for Traffic Control Strategy"
NCTRFPSCTH FSH 0418 "10.3.2 GPRS/EGPRS Traffic Control
Strategy"
NMO FSH 0720 "9.8.3.1 Network Operation Modes for
Paging"
NNSVCBLKR FSH 0720 "7.2.2.2 Control Procedures"
NNSVCRR FSH 0720 "7.2.2.2 Control Procedures"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

A30808-X3247-L24-5-7618 291
GPRS/EGPRS Global Description Information
System

Parameter Feature/CR Chapters

NNSVCTSTR FSH 0720 "7.2.2.2 Control Procedures"


"7.3.3 SGSN-BSS Flow Control"
NNSVCUBLR FSH 0720 "7.2.2.2 Control Procedures"
NRLCMAX FSH 0720 "9.9.3.1 Uplink TBF Using the Acknowledged
CR - X617 Mode"
"9.9.4.1 Acknowledged and Unacknowledged
Modes on Downlink TBFs"
NSEI FSH 0720 "7.2.1 Sub-Network Service: Frame Relay on
Gb Interface"
"7.2.1.1 Examples of Addressing"
"7.3.1 BSSGP Addressing: BSSGP Virtual
Connections (BVCs)"
NSVCI FSH 0720 "7.2.1 Sub-Network Service: Frame Relay on
Gb Interface"
"7.2.1.1 Examples of Addressing"
"7.2.2.2 Control Procedures"
NSVLI FSH 0720 "7.2.1 Sub-Network Service: Frame Relay on
Gb Interface"
"7.2.1.1 Examples of Addressing"
NTWCNDRXP FSH 0418 "10.3.1 Network Controlled Cell Reselection"
NTWCOR FSH 0720 "10.3.1 Network Controlled Cell Reselection"
NTWCREPPIDL FSH 0418 "10.3.1 Network Controlled Cell Reselection"
NTWCREPPTR FSH 0418 "10.3.1 Network Controlled Cell Reselection"
"10.3.1.1 Measurement Reporting"
NUA FSH 0720 "7.1 Physical Layer"
PCMECH FSH 0720 "10.4.2.2 Packet Transfer Mode: Measure-
ments for Power Control"
PCML FSH 0720 "7.1 Physical Layer"
PCUID (FLR object - FSH 0720 "7.1 Physical Layer"
In previous releases it
"7.2.1.1 Examples of Addressing"
was called PCUN)
PCUN FSH 0720 REMOVED IN BR6.0
(PTPPKF object)
PERSTLVPRI1 FSH 0420 "9.8.2.6 Uplink Access on PRACH (Access
Persistence Control)"
PERSTLVPRI2 FSH 0420 "9.8.2.6 Uplink Access on PRACH (Access
Persistence Control)"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

292 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Parameter Feature/CR Chapters

PERSTLVPRI3 FSH 0420 "9.8.2.6 Uplink Access on PRACH (Access


Persistence Control)"
PERSTLVPRI4 FSH 0420 "9.8.2.6 Uplink Access on PRACH (Access
Persistence Control)"
PKTMEASREPCNT FSH 0418 "10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
PKTNDEC FSH 0720 "9.9.3.1 Uplink TBF Using the Acknowledged
Mode"
"9.9.3.2 Uplink TBF Using the Unacknowl-
edged Mode"
PKTNINC FSH 0720 "9.9.3.1 Uplink TBF Using the Acknowledged
Mode"
"9.9.3.2 Uplink TBF Using the Unacknowl-
edged Mode"
PKTNMA FSH 0720 "9.9.3.1 Uplink TBF Using the Acknowledged
Mode"
"9.9.3.2 Uplink TBF Using the Unacknowl-
edged Mode"
PRPBCCH FSH 0720 "10.4.2.1 Packet Idle Mode: Measurements
for Power Control"
"10.4.3 BTS Output Power"
QSRHPRI CR - X482 "10.2.2 GSM-UMTS Re-selection Algorithm:
Packet Switched Case"
RAARET FSH 0720 "10.1.5 Abnormal Cell Re-selection"
RACODE FSH 0720 "9.2 Network Structure"
"10.1.4.3 Configuration of an Adjacent Cell
with GSUP= TRUE"
RACOL FSH 0720 "9.2 Network Structure"
"9.2 Network Structure"
RAENV FSH 0444 "10.5 Link Adaptation"
"10.5.1.3 GPRS: Link Adaptation Algorithm"
RARESH FSH 0720 "10.1.3 Cell Re-selection Algorithm"
REMAL FSH 0720 "7.1 Physical Layer"
SCHWEIPRI1 FSH 0550 "9.9.7.2 Scheduling Process"
"5.3.3.1 PCU Algorithm"
SCHWEIPRI2 FSH 0550 "9.9.7.2 Scheduling Process"
"5.3.3.1 PCU Algorithm"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

A30808-X3247-L24-5-7618 293
GPRS/EGPRS Global Description Information
System

Parameter Feature/CR Chapters

SCHWEIPRI3 FSH 0550 "9.9.7.2 Scheduling Process"


"5.3.3.1 PCU Algorithm"
SCHWEIPRI4 FSH 0550 "9.9.7.2 Scheduling Process"
"5.3.3.1 PCU Algorithm"
STGTTLLIINF FSH 0444 "10.5.1.2 “Quality Traps” Disadvantage"
"10.5.3 Selection of the Candidate Initial
Coding Scheme"
T1 FSH 0720 "7.3.1.1 BVC Procedures"
T2 FSH 0720 "7.3.1.1 BVC Procedures"
T3169 FSH 0720 "9.9.3.3 Anomalies During an Uplink TBF"
"9.9.3.4 Release of an Uplink TBF"
T3172 FSH 0720 "9.8.2.6 Uplink Access on PRACH (Access
Persistence Control)"
T3191 FSH 0720 "9.9.4.2 Release of a Downlink TBF"
T3192 FSH 0720 "9.9.4.2 Release of a Downlink TBF"
CR - X617
T3193 FSH 0720 "9.9.4.2 Release of a Downlink TBF"
CR - X617
T391 FSH 0720 "7.2.1.3 Procedures for PVCs"
TAVGT FSH 0720 "10.4.2.2 Packet Transfer Mode: Measure-
ments for Power Control"
TAVGW FSH 0720 "10.4.2.1 Packet Idle Mode: Measurements
for Power Control"
TCONG FSH 0720 "7.2.1.2 Frame Relay Structure"
TCONOFF FSH 0720 "7.2.1.2 Frame Relay Structure"
TDDARFCN FSH 0418 "10.2.3 Handling of UMTS Neighboring Cells"
TDDDIV FSH 0418 "10.2.3 Handling of UMTS Neighboring Cells"
TDDGQO FSH 0418 "10.2.2 GSM-UMTS Re-selection Algorithm:
Packet Switched Case"
TEMPCH FSH 0720 "5.3.1 Generalities about Resource Assign-
ments"
"5.3.3.1 PCU Algorithm"
"5.3.3.2 TDPC Algorithm"
TEMPPDT FSH 0429 "5.3.1 Generalities about Resource Assign-
ments"
TF1 FSH 0720 "7.3.3 SGSN-BSS Flow Control"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

294 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Parameter Feature/CR Chapters

THPROXT FSH 0720 This parameter was introduced for an older


CR - F287 SMG release, and it is no longer used starting
CR - X617 from BR6.0 release.
THSULBAL CR - X1495 "9.9.5 Notes About Concurrent TBFs"
TIMEDTBFREL CR - X617 "9.9.4.2 Release of a Downlink TBF"
TNSVCBLK FSH 0720 "7.2.2.2 Control Procedures"
TNSVCPTST FSH 0720 "7.2.2.2 Control Procedures"
"7.3.3 SGSN-BSS Flow Control"
TNSVCR FSH 0720 "7.2.2.2 Control Procedures"
TNSVCTST FSH 0720 "7.2.2.2 Control Procedures"
"7.3.3 SGSN-BSS Flow Control"
TRESEL FSH 0720 "10.1.5 Abnormal Cell Re-selection"
TRFPS FSH 0418 "10.3.2 GPRS/EGPRS Traffic Control
Strategy"
TRFPSCTRL CR - X1850 "10.3.1.2 Radio Link Network Controlled Cell
Reselection Algorithm"
TRFPSCTRLT FSH 0418 This parameter in BR7.0 release is not signifi-
cant.
TRXMD FSH 0420 "5.1.1 Enabling GPRS Service in the Cell"
"5.1.2 Enabling EGPRS Service in the Cell"
"5.1.3 Aspects Related to Carrier Configura-
tion"
"9.8.2.4 TBF Establishment for EDGE Mobile
Stations"
"5.2 Configuration of GPRS Channels in a
Cell"
TSULBAL CR - X1495 "9.9.5 Notes About Concurrent TBFs"
UPGRFREQ FSH 0516 "5.3.4.1 Upgrade of Radio Resources"

Tab. 11.1 GPRS/EGPRS Parameters Summary Table

Parameter Object Feature/CR Chapters

ASSLAPD SUBTSLB FSH 0419 "6.3.3 Configuration of the Abis Interface"


BSCT17 BSC ---- "8.2 PCU Overload Management"
BSCT18 BSC ---- "8.2 PCU Overload Management"
CELLGLID TGTFDD CR - X260 "10.2.3 Handling of UMTS Neighboring
Cells"

Tab. 11.2 Unspecific GPRS/EGPRS Parameters Involved in PS Services

A30808-X3247-L24-5-7618 295
GPRS/EGPRS Global Description Information
System

Parameter Object Feature/CR Chapters

CELLRESH BTS ---- "10.1.3 Cell Re-selection Algorithm"


CPOLICY BSC FSH 0457 "5.3.3 Management of Incoming
CR - F092 GPRS/EGPRS Requests"
CR - X1150
"5.3.3.1 PCU Algorithm"
"5.3.3.2 TDPC Algorithm"
"5.3.6.1 Pre-emption of PDCH Channels"
ENFOIAHO BSC FSH 0457 "5.3.6.3 Forced Intracell Handovers of
CR - F092 Already Established CS Calls"
DGRSTRGY BSC FSH 0457 "5.3.6 Waiting Queue Management"
CR - F092
"5.3.6.1 Pre-emption of PDCH Channels"
CR - X912
CR - X1150
FDDARFCN TGTFDD CR - X260 "10.2.3 Handling of UMTS Neighboring
Cells"
FDDDIV TGTFDD CR - X260 "10.2.3 Handling of UMTS Neighboring
Cells"
FDDQO BTS CR - X260 "10.2.1 GSM-UMTS Re-selection Algo-
rithm: Circuit Switched Case"
FDDQMI BTS CR - X260 "10.2.1 GSM-UMTS Re-selection Algo-
CR - X482 rithm: Circuit Switched Case"
"10.2.2 GSM-UMTS Re-selection Algo-
rithm: Packet Switched Case"
FDDSCRMC TGTFDD CR - X260 "10.2.3 Handling of UMTS Neighboring
Cells"
MSTXPMAXCH BTS ---- "10.1.2.1 GPRS/EGPRS Path Loss Crite-
rion (C1 Criterion)"
NBLKACGR BTS ---- "9.8.3.2 Discontinuous Reception"
NFRAMEPG BTS ---- "9.8.3.2 Discontinuous Reception"
NTWCARD BSC FSH 0397 "4.2.1 GPRS Channel Coding"
"6.1 Supported BSC Types"
"5.1.2 Enabling EGPRS Service in the
Cell"
QSRHI BTS CR - X260 "10.2.1 GSM-UMTS Re-selection Algo-
rithm: Circuit Switched Case"
RNCID TGTFDD CR - X260 "10.2.3 Handling of UMTS Neighboring
Cells"
RXLEVAMI BTS ---- "10.1.2.1 GPRS/EGPRS Path Loss Crite-
rion (C1 Criterion)"

Tab. 11.2 Unspecific GPRS/EGPRS Parameters Involved in PS Services

296 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

Parameter Object Feature/CR Chapters

SISGSNREL99 BSC FSH 0420 "7.3.3 SGSN-BSS Flow Control"


FSH 0514
"3.2 Network Architecture"
TDDQO BTS FSH 0418 "10.2.1 GSM-UMTS Re-selection Algo-
rithm: Circuit Switched Case"
TGTCELL ADJC FSH 1928 "10.1.4 Management of GPRS/EGPRS
CR - F119 Neighboring Cells"
"10.1.4.3 Configuration of an Adjacent
Cell with GSUP= TRUE"
TGTCELL ADJC3G CR - X260 "10.2.3 Handling of UMTS Neighboring
Cells"

Tab. 11.2 Unspecific GPRS/EGPRS Parameters Involved in PS Services

Object Feature Chapters

FRL FSH 0720 "7 Gb Interface"


NSVC FSH 0720 "7 Gb Interface"
PCMG FSH 0720 "7 Gb Interface"
PCU FSH 0720 "6 Hardware and Software Architecture"
PTPPKF FSH 0720 "4 Radio Interface Description"
TGTPTPPKF FSH 1928 "10.1.4 Management of GPRS/EGPRS
CR - F119 Neighboring Cells"

Tab. 11.3 Object Related to the GPRS/EGPRS Services only

Object Feature/CR Chapters

ADJC ---- "10.1 Cell Selection and Re-selection"


ADJC3G CR - X260 "10.2.3 Handling of UMTS Neighboring Cells"
SUBTSLB FSH 0419 "6.3.3 Configuration of the Abis Interface"
TGTBTS FSH 1928 "10.1.4 Management of GPRS/EGPRS Neighboring
CR - F119 Cells"
TGTFDD CR - X260 "10.2.3 Handling of UMTS Neighboring Cells"
TGTTDD FSH 0418 "10.2.3 Handling of UMTS Neighboring Cells"

Tab. 11.4 Unspecific GPRS/EGPRS Objects Involved in PS Services

A30808-X3247-L24-5-7618 297
GPRS/EGPRS Global Description Information
System

12 Abbreviations
AGCH Access Grant Channel
BCCH Broadcast Control Channel
BECN Backward Explicit Congestion Notification
BSC Base Station Controller
BSN Block Sequence Number
BSS Base Station Subsystem
BSSGP Base Station System GPRS Protocol
BVC BSSGP Virtual Connection
BVCI BSSGP Virtual Connection Identifier
CCCH Common Control Channel
CCU Channel Control Unit
CS Circuit Switched
DCE Data Circuit-terminating Equipment
DE Discard Eligibility Indicator
DLCI Data Link Connection Identifier
DRX Discontinous Reception
DTE Data Terminal Equipment
EGPRS Enhanced General Packet Radio Service
FDD Frequency Division Duplex
FECN Forward Explicit Congestion Notification
FR Frame Relay
FRL Frame Relay Link
GGSN Gateway GPRS Support Node
GMSK Gaussian Minimum Shift Keying
GPRS General Packet Radio Service
HA Horizontal Allocation
HCS Hierarchical Cell Structures
HLR Home Location Register
HSCSD High Speed Circuit Switched Data
HSN Hopping Sequence Number
IMSI International Mobile Subscriber Identity
IP Internet Protocol
IR Incremental Redundancy
LA Location Area
LAC Location Area Code
LAPD Link Access Procedure on the D-channel
LLC Logical Link Control
LMT Local Maintenance Terminal
MA Mobile Allocation
MAC Medium Access Control

298 A30808-X3247-L24-5-7618
Information GPRS/EGPRS Global Description
System

MAIO Mobile Allocation Index Offset


MCC Mobile Country Code
MCS Modulation and Coding Scheme
MNC Mobile Network Code
MOI Managed Object Instance
MS Mobile Station
MSC Mobile Switching Centre
NS SDU Network Service Service Data Unit
NS Network Service
NSEI Network Service Entity Identifier
NSVC Network Service Virtual Connection
NSVCI Network Service Virtual Connection Identi-
fier
NSVL Network Service Virtual Link
NSVLI Network Service Virtual Link Identifier
NUC Nailed Up Connection
O&M Operation and Maintenance
PAGCH Packet Acces Grant Channel
PBCCH Packet Broadcast Control Channel
PCCCH Packet Common Control Channel
PCH Paging Channel
PCU Packet Control Unit
PDCH Packet Data Channel
PDT Packet Data Terminal
PDTCH Packet Data Traffic Channel
PDU Packet Data Network
PDU Packet Data Unit
PLMN Public Land Mobile Network
PPCH Packet Paging Channel
PRACH Packet Random Acces Channel
PS Packet Switched
PSI Packet System Information
PSK Phase Shift Keying
PTCCH Packet Timing Advance Control Channel
PTM Point to Multipoint
P-TMSI Packet Temporary Mobile Subscriber Iden-
tity
PTP Point to Point
PTPPKF Point To Point Packet Function
PVC Permanent Virtual Circuit
QoS Quality of Service
RA Routing Area

A30808-X3247-L24-5-7618 299
GPRS/EGPRS Global Description Information
System

RAC Routing Area Code


RACH Random Access Channel
RAI Routing Area Identity
RAT Radio Access Technology
RC Radio Commander
RLC Radio Link Control
SGSN Serving GPRS Support Node
SMS Short Message Service
SNDCF SubNetwork Dependent Convergence
Protocol
SS7 Signalling System number 7
TAI Timing Advance Index
TBF Temporary Block Flow
TCH Traffic Channel
TDD Time Division Duplex
TDMA Time Division Multiple Access
TFI Temporary Flow Identity
TLLI Temporary Logical Link Identity ,
TSC Training Sequence Code
UE User Equipment
UMTS Universal Mobile Telecommunication
System
USF Uplink State Flag
VA Vertical Allocation
VLR Visitor Location Register

300 A30808-X3247-L24-5-7618

You might also like