Professional Documents
Culture Documents
1. Overview
2. Architecture
3. Channel Structure
6. Capacity Management
7. KPIs
2
Outline
3
HSDPA Basic Principles
4
1 - Shared Channel Transmission (1/2)
SF=2
TTI
Shared
channelization
codes
6
2 - Short 2 ms TTI (1/2)
2 ms
• The Transmission Time Interval becomes extremely short in HSDPA; 2 ms
compared to the 10 ms used by R99 high bit rate radio bearer.
• The HS channels are organised in sub-frame of 3 slots each; this means that
the slot time 2/3 ms/slot is the same as for R99 slots (10/15 ms/slot).
• The scheduling and the link adaptation algorithms work at this frequency!
Rel 5 (HS-DSCH)
2 ms
Earlier releases
10 ms
20 ms
40 ms
80 ms
7
2 - Short 2 ms TTI (2/2)
2 ms
• The shorter TTI allows:
– Reduced air-interface delay: this is required by the the TCP at high data
rates to Improved end-user performance
8
3 - HSDPA – Power Allocation
• HS-DSCH allocated power is decided by the RNC, prioritizing the DCH channel
• HS-DSCH adjusts the data rate to match the instantaneous radio conditions and the
available transmission power in the RBS
– No closed loop power control is specified for HS-DSCH, unlike the DCH channel
• The system adjusts the data rate by
– varying the effective code rate
– changing the modulation scheme
• This leads to a higher efficiency in the usage of power.
Power Power
3GPP Release 99 3GPP Release 5
Unused power
HS-DSCH (rate controlled)
• The target for the link adaptation is to select a TFRC (Transport Format and
Resource Combination) resulting in transmitting an as large transport block as
possible with a reasonable error probability.
Channel Condition
Coding
10
4 - Fast Link Adaptation (2/3)
feedback
High data rate
11
4 - Fast Link Adaptation (3/3)
feedback
High data rate
12
5 - Fast Channel-dependent Scheduling (1/2)
13
5 - Fast Channel-dependent Scheduling (1/2)
User 1
low data rate
User 2 Time
#1 #2 #1 #2 #1 #2 #1
14
6 - Fast Hybrid ARQ with Soft Combining (1/2)
• HSDPA introduces a new retransmission level “under” the RLC scheme in the
RNC.
• This new “level” allows rapid retransmissions of erroneous data:
– Hybrid ARQ protocol terminated in RBS
short RTT (typical example: 12 ms)
– Soft combining in UE of multiple transmission attempts
reduced error rates for retransmissions
P1,1 P1,2 P2,1 P2,2 P3,1
Transmitter
K
K
ACK
ACK
ACK
NAC
NAC
P1,2 P2,2
+ +
P1,1 P1,1 P2,1 P2,1 P3,1
Receiver
15
6 - Fast Hybrid ARQ with Soft Combining (2/2)
16
7 - UE capabilities
17
Throughput level: UE type cat 12
• There are several levels for throughput calculation: let’s clarify!
• The biggest MAC-HS transport block size is 3440 including HS header and padding bits:
X 10 = + + = 3440 bits
RLC SDU RLC MAC- HS SDU MAC- HS Padding
= 320 head = 16 head = 3360 head = 21 bits = 59
That means the DSCH max scheduled bit rate could be 1720 kb/s:
That is including headers, padding and every type of retransmission
This is the level used by the RBS counters and Couei!
This corresponds to a max RAB bit rate of 1600 kb/s =320*10/2
In reality considering at least the HS retransmissions at this level the maximum bit rate could not be higher
than 1600*0.9 =1440 bit/s
18
Agenda
1. Overview
2. Architecture
3. Channel Structure
6. Capacity Management
7. KPIs
19
Protocol stacks (1/4)
20
Protocol stacks (2/4)
21
Protocol stacks (3/4)
UE Uu RBS Iub SRNC Iu CN
RRC RRC
RLC RLC
MAC-d MAC-d GTP-U GTP-U
HS-DSCH HS-DSCH UDP/IP UDP/IP
FP FP
MAC-hs MAC-hs
AAL2 AAL2 AAL5 AAL5
ATM ATM ATM ATM
L1 L1 PHY PHY PHY PHY
22
Protocol stacks (4/4)
UE Uu RBS Iub SRNC Iu CN
TCP/IP TCP/IP
RLC RLC
MAC-d MAC-d GTP-U GTP-U
HS-DSCH HS-DSCH UDP/IP UDP/IP
FP FP
MAC-hs MAC-hs
AAL2 AAL2 AAL5 AAL5
ATM ATM ATM ATM
L1 L1 PHY PHY PHY PHY
24
HSDPA Basics: Node Impacts
• RBS
– New TX board in RBS
• RNC R99:
• Scheduling, Core
– No HW upgrades
• TF selection, Network
– Only SW!! • Link layer
RNC
– Setup of HS-DSCH/HS-SCCH retransmission
(ARQ)
HSDPA: Node B
• Scheduling,
• Link
Adaptation,
• Hybrid ARQ
25
Node Functionality
Iu
RNC Function:
RNC
Iur
RAB establishment & release
Channel switching
Mobility
Iub
Resource handling
Associated
Dedicated Capacity management
HS Channels
-DS
HS CH L2 (MAC-d)
HS -SCC
-DP H
CC
H UL HS control channel power control
26
Node Functionality
Iu
RBS Function:
RNC
Iur
L2 (MAC-hs)
Scheduling
27
Agenda
1. Overview
2. Architecture
3. Channel Structure
6. Capacity Management
7. KPIs
28
HSDPA new channels
HS
HS -DSC A-DCH
HS -SCCHH
-DP
CC
H
29
Channel Structure
Associated
Dedicated
HS Channels
-DS
CH
HS
HS -SCC
-DP H
CC
H
Control Channel
30
An overview of HS-DSCH and its associated channels
CN RNC RBS UE
RRC DCCH
For each DPCCH
user DCCH DCH
NAS DPCH DPDCH
DCCH DCH
NAS HS
- DPCCH
DCCH
HS-SCCH
Interactive PS RAB
DTCH
User 1
Interactive PS RAB
DTCH HS-PDSCH
User 2 HS -DSCH
Interactive PS RAB
User n DTCH
Iu Iub Uu
Radio Access Bearers: Logical Channels: Transport Channels: Physical Channels:
- Interactive -Dedicated Control Channel, DCCH -Dedicated Channel, DCH -Dedicated Physical Channel, DPCH
- Background -Dedicated Traffic Channel, DTCH -High-Speed Downlink Shared Channel, HS-DSCH -DPCCH, Dedicated Physical Control Channel
-DPDCH, Dedicated Physical Data Channel
-HS-DPCCH, HS-DSCH Dedicated Physical Control Channel
-HS-DSCH Shared Control Channel, HS-SCCH
-High Speed Physical Downlink Shared Channel, HS-PDSCH
31
HS-DSCH : High-Speed Dedicated Shared Channel
32
HS-SCCH: High-Speed Shared Control Channel
33
A-DCH: Associated Dedicated Channel
• One A-DCH pair is set up for every HSDPA user in connected state.
• It is used for control signaling (RRC and NAS) in UL and DL.
• It is a new Radio Bearer corresponding to a 3.4 kbps SRB in the DL (Sf
256).
• In the uplink A-DCH is also used as the channel for data transmission,
where the rate can be either 384 kbps or 64 kbps.
– The uplink data rate 384 kbps is selected as first priority and 64 kbps is used as a
fall back rate if the path loss is judged to be too large or 384 kbps radio bearer
setup fails for any reason (e.g. lack of radio or hardware resources).
• The uplink A-DCH channel also contains the High-Speed Dedicated
Physical Control Channel (HS-DPCCH), the new physical channel that
carries the L1 related signaling in UL.
34
HS-DPCCH: High-Speed Dedicated Physical Control
Channel
• It is used for transmitting the following information from UE to RBS:
– HARQ acknowledgement (1 bit coded in 10)
– Channel quality indicator (5 bits coded to 20 bits in 2 slots)
– channel quality measurements based on CPICH
– reporting rate is configurable through RRC/NBAP signaling
– information reflecting the instantaneous downlink radio channel conditions to assist the
RBS in the transport format selection (fast link adaptation) and the scheduling
– The HS-DPCCH has a spreading factor SFHS-DSCH = 256
• The A-DCH both UL and DL can be in soft/softer handover whilst the HS-
DPCCH can never be in soft handover (softer is possible).
• HS-DPCCH (UL) is transmitted within a dedicated channel. The main idea
is that it is power controlled from the other part of the A-DCH.
35
HSDPA Channel Operation
Physical Channels:
-Dedicated Physical Channel, DPCH
-DPCCH, Dedicated Physical Control Channel
-DPDCH, Dedicated Physical Data Channel
-HS-DPCCH, HS-DSCH Dedicated Physical Control Channel
-HS-DSCH Shared Control Channel, HS-SCCH
-High Speed Physical Downlink Shared Channel, HS-PDSCH
HS-DPCCH: CQI
HS-DPCCH: ACK/NACK
36
UserPlane: Overview
37
RAB/RB Combination – Overview
38
Agenda
1. Overview
2. Architecture
Accessibility:
3. Channel Structure Call setup phases
6. Capacity Management
7. KPIs
39
Radio network functions
Iu Iu
f1 f1 f1
f1 f1 f1
1
40
State Diagram – non HSDPA-P4
SRB CS Streaming
SP64
PS Streaming+Int 8/8
UDI (CS64)
Int. 64/128
SP0
Int. 64/64
UDI+Int. 8/8 New!
New!
SP0 not available
RAB Release
Int. FACH SRB
RAB Establishment RAB est on FACH
Channel Switching Note 1: It is possible to go to Idle from all states (Signaling connection release)
Note 2: QoS profiling on the PS Int RAB is handled by Channel Switching
Note 3: Same transitions is valid for PS Streaming 16/64 and PS Streaming 16/128
Note 4: RAB establishment on FACH depending on the setting for parameter PacketEstMode
41
State Diagram – only HSDPA
42
Idle mode & RRC Connection Establishment
f2 HS f2 HS f2 HS
Ec/No
HS+f1 f1 f1 f1 HS+ f1
f2
f1
UE UTRAN SGSN
SRB - DCH
0
RANAP: RAB Assignment Request
(establish PS Interactive/Background RAB)
2
Possible Inter frequency hard handover
3
Radio Bearer setup
44
Capability analysis
1
• At the reception of RANAP RAB Assignment Request, if:
– the present UE state is SRB-DCH and if the RAB mapping gives as result PS interactive
or PS background
– the “Access stratum release indicator” received from UE indicates Rel-5 or later
release,
– the “Physical channel capability” received from UE indicates that the UE supports FDD
HS-PDSCH (any HS-DSCH-physical-layer-category shall be supported),
– if the existing UE capability check for L2 are successful
UE UTRAN SGSN
SRB - DCH
46
Coverage relations 2
• The coverage relation is a unique uni-directional relation between two
cells, a source and a target cell.
• The purpose of the coverage relation is to give the operator a possibility to
distribute HSDPA downlink traffic among the cells of an RNC.
• The target cell covers almost the same area and can be assigned the
same frequency or different ones. Typically the cells will be co-located.
A coverage relation is
defined for a source cell
with the parameters (3GPP
R5 25.423):
hsPathLossThreshold
utranCellRef (the target cell)
coverageIndicator
47
Two Carriers scenario: IF HO 3
• In case the HSDPA is deployed on a second layer and the mobile access the
network from the first layer the step will be the following:
When the attempt on the AS cells fails, check the coverage relation of f2 HS
the best cell and its path loss f2 HS
f2 HS f2 HS
48 1st carrier
Interfrequency load distribution
49
UE RBS
RBS DRNC
DRNC SRNC
RB setup (1/2) 1. Admission request
• If the result from the “Serving 4. RNSAP: Radio Link Reconfiguration Prepare
HS-DSCH cell selection” gives
that a Serving HS-DSCH cell 4. Admission request
-> the “RB setup, SRB-DCH to PS 4. NBAP: Radio Link Reconfiguration Prepare
interactive (64 or 384)/HS - HS-
5. Allocate resources
DSCH transition” is performed 5. Allocate resources
50
RB setup (2/2) 4
• Different levels of Admission control runs in the RNC
– For the selected serving HS-DSCH cell, run Admission Control algorithm for the
A-DCH configuration and for HS-DSCH configuration (number of serving links).
– For the other cells within SRNC, run Admission Control algorithm for the A-DCH
configuration.
• The RANAP RAB Assignment Response is sent to the CN when the Radio
Bearer Setup Complete has been received.
• The handling of UL/DL user data on RLC level is done as in the R99 for
PS interactive RB.
51
UL: 64 or 384? 4
• Which UL A-DCH to set is decided during the AC phase.
• From an AC point of view, there are 2 guaranteed-hs service types:
– PS64/HS Interactive PS service with rate 64 kbps in uplink and HS-DSCH using
up to 5 HS-PDSCH codes in downlink.
– PS384/HS Interactive PS service with rate 384 kbps in uplink and HS-DSCH
using up to 5 HS-PDSCH codes in downlink.
• A part from the other AC check there are 2 special checks for the UL A-
DCH:
– Histogram Admission Policy: requests demanding spreading factor 4 in uplink
(PS384/HS radio connection type) are compared with sf4AdmUl.
– The path loss is checked in order to understand if a 384 UL bearer can be
sustained.
• If the 384 RB is denied (or is accepted but the RBS don’t find the
synchronization in the establishment phase) the connection is established
on the 64 RB.
52
Release of Iu-PS connection due to inactivity in HS-
DSCH state (1/2)
• When a user finishes its
transmission has to release the
resources.
• The procedure is really simple:
– An "HS-DSCH inactivity" timer is
started when there is no data to
transmit.
– When the timer expires a Iu
Release request is sent and the
resources are released.
• The value of the timer is a system
parameter hsdschInactivityTimer
53
Please note that..
54
Please note that.. Speech
Iu
RNC
call
Incoming CS call
• A critical issue for the HSDPA in P4 is the
management of the incoming call. Iub
1. Overview
2. Architecture
3. Channel Structure
7. KPIs
56
HSDPA Mobility: introduction (1/3)
• After the cell selection, the network has • 2 algorithms are interested in the
to guarantee the mobility of the HSDPA mobility of HSDPA users in connected
users mode:
– Serving HS-DSCH Cell Change
• Since No Soft/Softer HO exist for HS-
DSCH, there will be only one serving – A-DCH Soft/Softer HO
cell for the HS-DSCH. Iu Iu
RNC RNC
Iur
Iub
Iub Iub
Associated
Dedicated
HS Channels
-
HS DSCH
HS -SCC
-DP H
CC
H
57
HSDPA Mobility: introduction (2/3)
58
HSDPA Mobility: introduction (3/3)
f2 HS 2 layers
1 layer f2 HS f2 HS
f f2 HS f1 f1 f1
HS HS f f
f1 f1 f1
59
Measurement reporting (1/2)
60
Measurement reporting
• Note that the RNC can configure more than 1 measurement report for the
same event.
61
Mobility Example
2
1 3
Active set handling:
(Max active set = 3) 0
cell 1
cell 2
event 1a: cell 3
Add cell 2, to the AS -5
Add cell 3, to the AS
Ec/N0 [dB]
Delete cell2
-15
event 1d:
Change of best cell, to cell 3
-20
0 5 10 15
time [s]
62
Measurement reporting for HSDPA
63
A-DCH handover
• In the previous slide it is stated that “there will be only one serving cell for
the HS-DSCH”. This does not mean that the UE is connected to only one
cell.
• For what concerns the A-DCH they continue to perfrom soft and softer
hand-over as in normal R99 case.
• Note that HS-DPCCH can be only in softer HO.
• In the example supposing the best server does not change...
A-DCH
HS-DSCH
R99 HS HS
R99 R99 HS
HS HS
HS
HS HS
64
Serving HSDPA Cell Change (1/2)
65
Serving HSDPA Cell Change (2/2)
66
Soft/softer HO for A-DCH and cell change for HSDPA
channels
P_CPICH 2 hysteresis _ 1d hs
A-DCH on cell 2
only
time
Reporting Reporting Reporting
event 1a event 1d-hs event 1b
HSDPA channels cell
A-DCH in SHO with cell 1 and 2 change from cell 1 to cell 2
67
Serving HSDPA Cell Change
• If a suitable HS-DSCH cell can not be found within the current RNS a RRC
connection release is triggered.
• After this a new cell selection can follow:
– Another connection establishment with a new Cell selection
– normal connection establishment on R99
– A connection establishment on GSM
Ex3
f2 HSf2 HS f2 HS
Ex1
f HS R99 R99 R99
f2 HS 2 f2 HS
R99 R99 R99
Cell Selection
f HS f HS Ex 4
f1 HS 2 f1 HS f1 HS 2 f1 HS f2 HSf2 HS f2 HS
Cell Selection Ex 5
f HS f HS f HS
R99 2
HS R99
HS R99 2
HS HS R99 2
HS R99
HS R99
R99 R99 R99
68
Radio Connection Supervision
69
Agenda
1. Overview
2. Architecture
3. Channel Structure
6. Capacity Management
7. KPIs
70
“Carrier” mobility
• A user conected with a UMTS network on a certain carrier can move out of
the its layer coverage.
• There are 2 mechanisms to avoid the drop, at least for some services:
– Inter-RAT Handover
– Inter-Frequency Handover HS
-S C
CH
f2 HS
f2 HS f2 HS
UE moves to poor
1 carrier coverage area
and reports to UTRAN
UTRAN
UTRAN commands
CM measurements 2
• Main steps
– When the quality of the connection overcomes a certain thresholds (event 2d or
6a) UE reports a Measurement report and the network orders the UE to activate
the CM and check the other layer.
– If the connection quality further degrades and the other layer has a quality high
enough the HO is triggered (events 3a and 2b).
– If the connection quality turns out to be good, the UE signals it to the network and
the mobile stops the CM (events 2f or 6b).
Layer Quality
Threshold 1
Threshold 2
Threshold 3
time
CM L1
The problem is that there are several thresholds
HOCompressed Mode L2
73
IRAT & IFHO procedures
• At the same time, the cell in the target layer should have the quality good
enough. That means:
– For GSM: the quality of the measured GSM cells is above a gsmThresh3a.
– For the second UMTS layer: the measured best cell on unused frequency is
above both the thresholds nonUsedFreqThresh4_2bEcno and
nonUsedFreqThresh4_2bRscp
• Both of the HO are hard HO:
– This means that there will be small interruptions in the data flow to and from the
UE.
74
CPM start and HO trigger
Target
CPICH
Ec/No Layer
CPICH
RSCP
Start
CPM
UE TX Start
CPM
Initial power Other layer
Cell good enough
HO Start HO
trigg CPM trigg
HO
trigg
75
IRAT HO and Cell Change
76
IF or IRAT?
• A decision has to be made to evaluate either Inter-Frequency
handover or Inter-RAT Handover/Cell Change. This decision is based
on parameters on RNC level, cell level, and UeRc state.
Where?
– Inter-Frequency handover is only attempted if
C_IfHoAllowed is set to Allowed for the current UeRc state,
and FddIfHoSupp (RNC) is set to On .
– Inter-RAT handover is only attempted if C_GsmHoAllowed
is set to Allowed for the current UeRc state, and f2 HS
FddGsmHoSupp (RNC) is set to On.
– If both the conditions are verified the decision is based on a
configurable parameter, hoType (cell), defined per cell ?
(IFHO preferred, GSM preferred, None). f1
f1 ?
• Hence, for a certain cell only one of the 2 Handover types will
be allowed.
GSM
GSM
77
CIPICH dimensioning
• Dimensioning example
79
Impact on coverage
CM start Ec/No
Area
without CM
WCDMA RBS
HO Ec/N0
CM start RSCP
HO RSCP
80
Notes for HSDPA
• Anyway:
– they can experiment it when in connected mode with a R99 RAB or in other dedicated
connection.
– They can impact on other users behavior
81
IRAT Cell Reselection
Overall description of thresholds
Ec/No>qQualMin
WCDMA acceptable area
RSCP>qRxLevMin +P WCDMA->GSM normal
reselection area
GSM coverage
82
Agenda
1. Overview
2. Architecture
3. Channel Structure
6. Capacity Management
7. KPIs
83
Capacity Management (Overview)
84
Dedicated Resources
• RF POWER:
the total transmitted carrier power is constantly monitored by the algorithm. When the value exceeds some
configurable thresholds the admission/congestion take decisions for guaranteed and non-guaranteed service
class connections
• CODE
– Code Usage: the total number of codes is monitored.
– Code Hystogram: the number of codes used for each SF are monitored. The max number of code for
each SF is configurable. A control is also done on the maximum number of compressed mode connections
85
Admission Control Algorithm
86
Congestion Control Algorithm
Default values:
pwrOffset = +5% 90%
pwrHyst = 600 ms
87
Traffic Algorithms
88
HSDPA Monitored Resource Handling
– The monitored power in the RNC keeps track only of the usage of total non-HS
downlink transmitted carrier power.
– The reports of the power measurements are adapted to the capability of a cell:
– HSDPA capable ‘Transmitted carrier power of all codes not used for HS-PDSCH or
HS-SCCH transmission’
– HSDPA not capable ‘Transmitted carrier power’
– The measurement of code tree utilization considers the codes allocated for HS-
PDSCH and HS-SCCH channels.
89
HSDPA Admission Control Algorithm
f2 HS
f HS f2 HS1
R99 R99 R99
90
HSDPA Admission Control Algorithm
– HS-DSCH:
– Number of HS-serving links (for RAB set up only)
91
HSDPA - Code Control
SF
1
16
Common
Dedicated HS-PDSCH
Channels;
Channels (default)
HS-SCCH
93
HSDPA –UL Histogram AC
• The operator can set a limit for the ”guaranteed-hs” admission requests
demanding spreading factor 4 in uplink that can be accepted (in cells where the
PS384/HS is activated)
• This policy allows the operator to disable the PS384/HS feature on a cell basis
94
HSDPA DL Power Admission
guaranteed-hs / non-handover
• Highest priority for ”guaranteed-hs”
service class in admission decisions
non-guaranteed / handover
guaranteed / non-handover
guaranteed-hs / handover
to enable HSDPA users to use the
guaranteed / handover
excess power in high loaded (non-
congested) cells
non-guaranteed /
non-handover
• Soft congestion is not affected by
the introduction of HSDPA
Power
Reject
Reject
Reject
pwrAdm + pwrAdmOffset + pwrOffset
Admission Granted
Admission Granted
Admission Granted
pwrAdm – beMarginPwrAdm
Admission
Granted
95
HSDPA – “Link” Power Admission
Common channels
96
HSDPA - Congestion Control
Power
97
Agenda
1. Overview
2. Architecture
3. Channel Structure
6. Capacity Management
7. KPIs
98
Load Sharing (1/2)
Both load-sharing features redirect calls during the connection setup phase:
RRC connection setup for IF load Sharing
RAB setup for Directed retry
Both IF HO and Directed retry will be present at the same time but IF will act first!
99
Load Sharing (2/2)
• For load sharing purposes, cell load is defined as
Cell Load = Tx_power/PwrAdm
the ratio between the downlink transmitted carrier
power and the admission limit, as given by the cell
parameter pwrAdm. IF
Cell Load
• The load sharing algorithm acts only when the load 50%
of a cell overcomes a certain threshold:
– For IF, the threshold is set to 50%
– For directed retry both the threshold and the percentage
of users to be redirected can be tuned
Time
DR
Cell Load
thr
Time
• Speech call (without packet connection) is the only service that is targeted since
it is also the only one that is safe to divert to GSM
• Directed Retry is performed during the RAB establishment procedure;
– the first request will be rejected with cause "Directed retry“
– a request is made to the core network to relocate the UE to a specific GSM cell,
using the Inter-RAT handover procedure.
• This handover is a blind HO since the target cell is chosen not based on UE
measurements. Therefore, the target cell must be co-located with the WCDMA
cell.
There are 2 control parameters:
DR
Cell Load
loadSharingGsmThreshold
specifies the minimum cell load at which off-loading to loadSharingGsmThreshold
GSM begins.
loadSharingGsmFraction specifies
the percentage of Directed Retry candidates to be diverted
to GSM
Time
101
Inter-Frequency Load Sharing (1/2)
1 The mobile starts a RRC connection establishment procedure ( NO distinction in RRC cause is made)
2 If the cell load is higher than 50%, the load of the co-located load-sharing neighbor is compared with
the accessed cell and the least loaded cell is chosen as target.
3 If the target cell is less loaded, the UE will not be instructed directly to go to the target cell but it will
be told to scan for a suitable cell in the frequency of the target cell, by sending an RRC Connection
Reject message.
2
3 Load >=<
UMTS L2
Load
Cell 2 ? 50%
UMTS L1 1
Cell 1
GSM
102
IF Load Sharing
L[2]
L[2] == 33%
33%----L[1]
L[1]= =60%
60%
Layer 2 Second Carrier L[1]
% ->20%
L[1]
> 50
L[1] < L[2]>L[2] R[2],
Select
Select
Compare
-> Don’t Second
the load Carrier
do anything with the
33% Power/
Layer 1 33% candidate
load sharing
100% pwrAdm
Free Resource = R[2]
Load
loadSharingThreshold
First Carrier (20%)
>=< Free Resource
Load
Free Resource
33%
60% Power/
?
Cell 2
Cell 1 DL power in use 100%
pwrAdm
103
HSDPA – IF Load Sharing
loadSharingMargin is a cell-specific parameter that specifies the amount of resource excluded from load-
sharing use (as a percentage of pwrAdm).
When loadSharingMargin is greater than 0, the cell appears to be more loaded than it really is, resulting
in more traffic being directed away from it
loadSharingThreshold
>=<
Load
This parameter gives the operator the possibility to reserve a higher priority to the HSDPA users on
the second carriers (in case this is deployed and HSDPA is introduced there)
104
Inter-Frequency Load Sharing
• Apart from the load sharing algorithm, other aspects have to be managed
and tuned with the introduction of the second layer:
– Mobility:
– IF HO procedure is more critical compared to a normal SHO and has to be
verified and tuned.
– The compressed mode activity increases in the border cells.
– Accessibility:
– The first IF Load Sharing will increase the call set up time.
– Terminal equipment limitations:
– At them moment, there are several terminal types not fully supporting the
features to manage a second layer. In particular several models are not IF
HO capable.
105
Agenda
1. Overview
2. Architecture
3. Channel Structure
6. Capacity Management
7. KPIs
106
Coverage vs. traffic load
UL DL
high high
The DL coverage (considering the Load
Ec/N0 of the CPICH) in particular Loa
WCDMA RBS
decreases with the DL total power d
(hence with the load).
CPICH_Ec/N0 in a point: UL low DL low
load load
Ec/N0Cpich = RSCP/RSSI
= Pcpich/( (PtotIntra + Ptotinter + Noise)
(*P = received power)
107
DL Problem CPICH Ec/N0 triggering
1. Overview
2. Architecture
3. Channel Structure
6. Capacity Management
6. KPIs
109
Accessibility (CSSR) KPI
CSSR is currently calculated by two factors:
RRC Establishment Success Rate = RRC_Success / RRC_Attempts *
RAB Establishment Success Rate = RAB_Success / RAB_Attempts
The IF Load Sharing feature impacts on the RRC Establishment Success Rate
since several RRC Connection Attempts are rejected to be redirected towards the
other frequency. So it is expected that, in case of IFLS activated, the number of
RRC Connection Success will be reduced because of the Load Sharing Reject
events. To take into account this fact the LoadSharingRejects must be subtracted
from the total number of RRC_Attempts:
pmTotNoRrcConnectReqPsSucc
RRC _ Succ _ PS
pmTotNoRrcConnectReqPs
pmTotNoRrcConnectReqPs - pmNoLoadSharingRrcConn
pmTotNoRrcConnectReq
The basic idea is to calculate the load sharing reject for PS and CS by a
wheight factor given by the fraction of the RRC_CS (or PS) respect to the total
number of RRCs. In case of RRC Succ general the formula is muche more
simple instead....
pmTotNoRrc ConnectReqSucc
RRC _ Succ _ general
pmTotNoRrc ConnectReq - pmNoLoadSharingRrcConn
111
Load Sharing Impact on Accessibility (CSSR) KPI - Results
RRC Connection Establishment Success Rate CS and PS in case of Load Sharing
10000 120%
9000
100%
8000
7000
80%
6000
The RRC Succ estimation CS and PS are LoadSharingReject
disturbed by the Load Sharing rejects.
5000 60%
We registered strong fluctuations of RRC_SUC_CS_LS
RRC_SUC_PS_LS
4000
values expecially in case of high IFLS
activity....
40%
3000
2000
20%
1000
0 0%
Days
112
Load Sharing Impact on Accessibility (CSSR) KPI - Results
RRC Connection Establishment Success Rate General in case of Load Sharing
10000 120.0%
9000
100.0%
8000
7000
80.0%
4000
40.0%
3000
2000
20.0%
1000
0 0.0%
Days
113
Load Sharing Impact on Accessibility (CSSR) KPI - Results
RRC Connection Estab. Succ. Rate comparison (General, CS, PS) including Load Sharing
110.0%
>100% values
105.0%
100.0%
RRC_SUC_LS
RRC_SUC_CS_LS
95.0%
Here it is quite evident the noise introduced by
RRC_SUC_PS_LS
the load sharing in CS and PS RRC Succ
estimations.
90.0%
85.0%
Strong KPI
deterioration
80.0%
Days
114
Please consider that the values are calculated on daily base. So they
should be quite stable
Conclusions (1/2)
When Load sharing is introduce the accessibility formulae should be updated to
take into consideration the RRC Connection Attempts rejected to be redirected
towards the other frequency ().
While no problem should exist for the tot accessibility formula:
pmTotNoRrc ConnectReqSucc
RRC _ Succ _ general
pmTotNoRrc ConnectReq - pmNoLoadSharingRrcConn
pmTotNoRrcConnectReqPsSucc
RRC _ Succ _ PS
pmTotNoRrcConnectReqPs
pmTotNoRrcConnectReqPs - pmNoLoadSharingRrcConn
pmTotNoRrcConnectReq
However the estimation given by this KPI is not extremely stable/reliable and
accurate (the average error seems to be acceptable compared to the error that
affects the other formulae but the fluctuation are high).
115
Conclusions
An alternative suggestion could be to use the following formula for CSSR
The estimation given by this KPI is much more stable/reliable and the average
error seems to be acceptable compared to the error that affects the other
formulae.
The main drawback of this solution is that the Global RRC Succ is often a little bit
worse compared with the real CS and PS values.
(This is probably related to different radio environment: i.e. the major part
of RRC Connections are established for registration purpose, when the
UE is entering back to 3G coverage; those radio procedures often occur
at cell coverage borders and so are affected by a worse performance.)
116
Agenda
1. Overview
2. Architecture
3. Channel Structure
6. KPIs
117
RBS Counters KPIs
• This appendix reports the main KPIs that could be used for HS testing
divided by:
– Throughput
– Scheduling Ratio and Transmission efficiency
– CQI/ACK/NACK
– Power
– RBS RSSI
118
Throughput Counters
• Counters:
– pmSumAckedBits: the number of Media Access Control high-speed (MAC-hs) bits received and
acknowledged by the UE.
– pmSumTransmittedBits: the number of transmitted bits at MAC-hs, level including retransmissions
– pmSumNonEmptyUserBuffers: The number of user buffers containing high-speed data.
– pmNoActiveSubFrames: the number of subframes containing high-speed data transmitted by the
RBS.
– pmNoInactiveRequiredSubFrames: the number of empty subframes transmitted even though data is
scheduled for priority queue.
119
Throughput KPIs
120
Throughput KPIs Values – Live Network
Average
RBS DSCH UE DSCH UE DSCH Cell DSCH Cell
RBS Number
RBS Name Sector Throughput Throughput Throughput Throughput
Sector Of UE In
Carrier NET GROSS NET GROSS
a Queue
RBS32601-01 3 1 1 92.32 148.32 194.89 313.11
RBS05314-01 2 2 1 164.11 221.32 265.44 357.98
RBS37058-01 1 2 1 183.97 202.45 292.65 322.04
RBS00003-01 3 2 1 218.77 259.69 218.77 259.69
RBS34645-01 2 1 1 242.96 350.73 242.96 350.73
RBS01281-01 3 1 1 248.4 405.3 248.4 405.3
RBS37058-01 2 2 1 324.44 405.58 324.79 406.02
RBS01780-01 1 1 1 352.6 566.15 491.79 789.63
RBS00357-01 2 1 1 389 518.5 389 518.5
RBS00429-01 1 2 1 399.94 542.16 399.94 542.16
RBS37242-01 3 1 1 424.5 500.75 424.5 500.75
RBS34375-01 2 1 1.01 449.73 749.43 461.94 769.76
RBS01263-01 1 1 1 456.78 565.98 456.78 565.98
RBS23074-01 3 2 1 509.66 552.27 509.66 552.27
RBS23066-01 2 1 1 518.67 518.67 518.67 518.67
RBS04995-01 1 2 1 556.41 648.95 557.03 649.67
RBS00370-01 1 1 1 632.88 838.31 632.88 838.31
RBS01546-01 2 1 1 652.58 878.57 751.29 1,011.47
RBS01281-01 1 1 1 662.62 1,142.64 662.62 1,142.64
RBS34375-01 3 1 1 710.73 1,071.00 710.73 1,071.00
RBS23025-01 3 1 1.01 734.94 901.09 828.62 1,015.96
RBS37058-01 3 2 1 764.56 995.06 817.15 1,063.51
RBS00353-01 1 1 1 770.18 1,170.15 775.24 1,177.84
RBS34645-01 1 1 1 828.63 1,319.63 933.18 1,486.14
RBS23025-01 2 1 1 881.87 1,205.44 913.16 1,248.20
RBS05314-01 1 2 1 1,183.36 1,390.58 1,193.88 1,402.94
RBS00429-01 2 2 1 1,200.85 1,605.63 1,200.85 1,605.63
RBS01164-01 3 2 1 1,285.29 1,503.11 1,285.65 1,503.53
RBS01164-01 2 2 1 1,318.85 1,537.49 1,354.04 1,578.52
121
Transmission efficiency KPIs
• Scheduling Ratio = This KPI simply highlights the percentage of time the HS-DSCH is used
Sum(pmNoActiveSubFrame)*0.002/ (ROP period)
• Transmission ratio = This highlight the percentage of time there is something to transmit. This is a good index on
how efficiently the application level can exploit DSCH capabilities
(Sum(pmNoActiveSubFrame)+Sum(pmNoInactiveRequiredSubFrame))*0.002 /
(ROP period)
122
Transmission Efficiency (1/2)
• The first index to look at when talking about transmission efficiency is the ratio
between the used TTI (the one where something is transmitted) and the total
#TTI in the test period
124
CQI/ACK/NACK counters
125
CQI/ACK/NACK KPIs
• CQI specific
– Av. Reported
– Av. USed CQI
– Delta CQI = Difference between the 2 averaged values above. This is an index of how much “CQI adjustment” acts.
– CQIequalTo0 = Count(CQIreported=0)/Count(CQIreported). This is the main reason of MAC inefficiency
– InvalidCQI = invalideCQI/Count(CQIreported). Not clear what invlid means
(pmNackReceived+pmAckReceived)/ActiveFrame
126
HS scheduling The BLER is the ratio between NACK
and (ACK + NACK)
127
Integrity KPIs Values– Live Network
RBS MAC
RBS RBS Sector % HS- % Rtx Transmission
Name Sector Carrier BLER Overhead Efficiency
RBS23066-01 2 1 0.00% 0.00% 100.00%
RBS37058-01 1 2 4.14% 9.13% 62.87%
RBS37242-01 3 1 8.85% 15.23% 100.00%
RBS23074-01 3 2 9.43% 7.72% 100.00%
RBS05314-01 2 2 10.01% 25.85% 61.83%
RBS04995-01 1 2 11.74% 14.26% 99.89%
RBS01164-01 2 2 12.03% 14.22% 97.41%
RBS05314-01 1 2 12.99% 14.90% 99.12%
RBS01164-01 3 2 13.28% 14.49% 99.97%
RBS01263-01 1 1 13.64% 19.29% 100.00%
RBS23025-01 3 1 14.18% 18.44% 89.32%
RBS37058-01 2 2 14.34% 20.01% 99.89%
RBS00003-01 3 2 15.38% 15.76% 100.00%
RBS37058-01 3 2 18.06% 23.16% 93.56%
RBS34375-01 2 1 18.24% 39.99% 98.14%
RBS01546-01 2 1 18.33% 25.72% 87.02%
RBS00370-01 1 1 18.44% 24.51% 100.00%
RBS00429-01 1 2 19.62% 26.23% 100.00%
RBS32601-01 3 1 21.05% 37.76% 47.37%
RBS23025-01 2 1 22.70% 26.84% 96.60%
RBS34375-01 3 1 23.36% 33.64% 100.00%
RBS01780-01 1 1 23.64% 37.72% 71.70%
RBS00429-01 2 2 24.27% 25.21% 100.00%
RBS00357-01 2 1 25.00% 24.98% 100.00%
RBS34645-01 1 1 27.50% 37.21% 88.80%
RBS00353-01 1 1 28.36% 34.18% 99.35%
RBS34645-01 2 1 30.77% 30.73% 100.00%
RBS01281-01 3 1 33.68% 38.71% 100.00%
RBS01281-01 1 1 38.09% 42.01% 100.00%
128
RBS power
• Notes:
– Every 100 ms the transmitted carrier power for all non high-speed codes in the
cell are sampled. The problem is that there are not necessary data enough to
transmit in every slot: hence some kind of “normalization” should be investigated.
129
Tx Power examples
(almost 100% of Scheduling Ratio)
0.5
TxCarrierPower
0.4 R99Power
0.3
0.2
0.1
• Note that the HS power is an estimation of the power transmitted for the HS on average
during the test period but it is NOT an estimate of the power that HS required in the cell!
– The activity of the HS (when we transmit something) is still low (80%).
– Maybe a more interesting KPI could be:
– TotalHS_power/SchedulingRatio*100
= 100*3.22/81.3 = 3.96 W
• It is interesting to notice anyway that even with a single user the MaxTxPower is reached.
131
RBS list of counters
• pmTransmittedCarrierPowerNonHs: The transmitted carrier power for all non high-speed codes in the cell.
• pmTransmittedCarrierPower: the transmitted carrier power measured at the TX reference point every 4 seconds
• pmNoActiveSubFrames: the number of subframes containing high-speed data transmitted by the RBS.
• pmNoInactiveRequiredSubFrames: the number of empty subframes transmitted even though data is scheduled for priority queue.
• pmSumNonEmptyUserBuffers: The number of user buffers containing high-speed data.
• pmSumAckedBits:the number of Media Access Control high-speed (MAC-hs) bits received and acknowledged by the UE.
• pmSumTransmittedBits: Description The number of transmitted bits at MAC-hs, level including retransmissions
• pmReportedCqi: the Channel Quality Indicators (CQI) reported by the UE in the cell and received by the RBS.
• pmUsedCqi: the CQI, used by the RBS for scheduling the priority queue for the HS-DSCH. pmAckReceived: The number of
Acknowledgements (ACK) that the RBS receives from the User Equipment (UE) over the High-Speed Downlink Shared Channel (HS-
DSCH).
• pmNackReceived: The number of Negative-Acknowledgements (NACK) that the RBS receives from the User Equipment (UE) over the
High-Speed Downlink Shared Channel (HS-DSCH). pmAverageRssi: The average Received Signal Strength Indication (RSSI).
132
Agenda
1. Overview
2. Architecture
3. Channel Structure
6. KPIs
133
Cell Availability (1)
pmHsDowntimeMan
Av_Man_HS( Hsdsch) * 100
24 * 3600
• The length of time in seconds that a cell is available for Packet Interactive HS service is defined as cell HS availability. in the example, the cell HS availability during
24 hour period is reported.
134
Accessibility
• The newThe
andnumber
existingofcell countersRAB
attempted used in the PS RABfor
establishments establishment
PS Interactive RAB
procedure are given
mapped in the following
on HS-DSCH list: for the selected Serving HS-DSCH cell
(stepped
at RAB establishment and before possible Inter-Frequency HO).
pmNoHsHardHoReturnOldChTarget NumberNew
of failed Hard HO for serving HS-
DSCH cell selection and UE
135
connection maintained (in the target cell).
Accessibility/ IF counters
136
Accessibility
pmNoRabEstablishSuccessPacketInteractive
100 *
(pmNoRabEstablishAttemptPacketInteractive –
pmNoOutgoingHsHardHoAttempt + pmNoIncomingHsHardHoAttempt
+pmNoHsHardHoReturnOldChSource- pmNoHsHardHoReturnOldChTarget )
pmNoRabEstablishSuccessPacketInteractiveHs
100 *
pmNoRabEstablishAttemptPacketInteractiveHs
137
InterFrequency Handover
• The new and existing cell counters used for Retainability are given in the
following list:
pmNoSystemRabReleasePacket Existing
Number of successful normal releases of packet RABs
pmNoNormalRabReleasePacket mapped on HS-DSCH
Existing in the Serving HS-DSCH cell.
pmChSwitchFachIdle Existing
The number of signalling connection releases
pmNoTpSwitchSp64Speech Existing
triggered for PS Interactive RAB mapped on HS-
pmNoSystemRbReleaseHs
DSCH due to New
inactivity (Channel Switching
Evaluation algorithms request the execution of a
pmNoNormalRbReleaseHs switch to idle).
New
The counter is stepped at the reception of RANAP Iu Release
pmInactivityHsIdle Command fromNew CN, for HS channel cell or RANAP RAB
assignment Request (when the RAB is released) and the RANAP
cause is User Inactivity.
139
Retainability
pmNoSystemRbReleaseHs
100 *
pmNoSystemRbReleaseHs + pmNoNormalRbReleaseHs
( pmNoSystemRabReleasePacket )
100 *
pmNoSystemRabReleasePacket + pmNoNormalRabReleasePacket
140
System Utilization
• It is possible to measure HS A-DCH utilisation in terms of code usage and average number of users per cell.
• Two new set of counters shall be implemented. The first set is used to observe the HS A-DCH code utilisation per cell. The KPI indicates the the total number of A-DCH
radio bearers established in a cell.
– ( pmSumPsHsAdchRabEstablish /pmSamplePsHsAdchRabEstablish )
• The second set is used to observe the average number of users per cell (hence the number of HS users), which is done by looking only at the best cell:
– (pmSumBestPsHsAdchRabEstablish /pmSampleBestPsHsAdchRabEstablish )
141
Throughput
142
Throughput
Retrans Rate
Object Name
PS Int HS
RNCCN1 99.63%
RNCKS1 99.49%
RNCKS2 100.00%
RNCNY1 98.03%
144
Mobility
pmNoHsCcSuccess New
pmNoHsCcAttemptt New
pmHsCcSuccess
PS_M_HSCC_S *100
pmHsCcAttempt
145
Admission & Congestion Control
146