You are on page 1of 132

13.07.

2017

LTE Drive Test & Network Analysis

www.piworks.net
1.1 Post Processing Overview

www.piworks.net
Actix Overview

www.piworks.net
How to Load Logfile’s

www.piworks.net
What is Super Stream ?

Super
Logfile 1 Stream
.
.
.
Logfile N

A superstream is a new stream created by merging two or more sets of data into one
large data set. Beware that the size can be huge (since the superstream contains all
attributes, even those that we might not need in our analysis).

www.piworks.net
How to Create Superstream

www.piworks.net
Display all required information (1)

On the Superstream Attribute,


breakdown the information that we
want to show into (Map, Chart or
Table)

www.piworks.net
Display all required information (2)

6 All Detail Information

www.piworks.net
Display all required information (3)

RSRP Level

www.piworks.net
Display all required information (4)

Layer 3 Message

www.piworks.net
Complete Analysis Template

www.piworks.net
Limitation

• Actix suitable for TEMS, NEMO and XCAL Logfile.


• For HUAWEI Genex Probe only suitable by using Genex
Assistant.

www.piworks.net
1.2 Drivetest Reporting on Accessibility Performance

www.piworks.net
Accessibility Performance

Attach Full
3 Layer
message

www.piworks.net
Accessibility Performance (RACH)

RACH Fail
when SINR
Level very
poor

No KPI for RACH


since Random
Nature of RACH

www.piworks.net
Accessibility Performance (RRC)

RRC SSR[%]=
RRCConnectionSetupOK/
RRRCConnectionAttempt*
100%

RRC Connection Setup, not


No RRC Setup relevant with the
Failure in this report, RRCConSetupComplete
message
RRC SSR 100%

www.piworks.net
Conclusion of Accessibility Performance
• This method should be taken from short call Logfile.

• 3 Layer message available for analysis.

RRC Attempt

RRC Success
RRC Fail

www.piworks.net
1.2 Drivetest Reporting on Mobility Performance

www.piworks.net
Measurement Event

www.piworks.net
Parameter Configuration (Intra-Frequency Handover)
Source Target
UE
eNB eNB
Handover-related parameters are used to control the reporting time and 1.A3 Measurement Control
difficulty level of a handover in the measurement report. For details about
handover parameter configuration, see Measure RSRP/RSRQ

eRAN3.0 Handover-related MML Command Configuration Guide


•Intra-frequency handover triggering process (event A3) A3 EventTriger

Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off A3 Measurement


➢Mn indicates the measured Reference Signal Received Power (RSRP)
2.
Reports

and Reference Signal Received Quality (RSRQ) of neighboring cells. Handover preparation
➢Ofn indicates the frequency offset. 3
RRC Conn. Reconf. incl.

➢Ocn indicates the neighboring-cell offset to be configured through the


mobilityControlinformation

neighboring relationship.
Random access procedure
➢Ms indicates the measured RSRP and RSRQ of the serving cell.
➢Ofs indicates the offset of the serving frequency. 4 RRC Conn. Reconf. Complete

➢Ocs indicates the offset of the serving cell.


➢Hys indicates the hysteresis, closely related to the service feature and
the mobility speed.
Reducing the probability of ping-pong effect.
➢Off a3-Offset
➢Mn, Ms is in units of dBm or dB. Other parameters are in units of dB.

www.piworks.net
Parameter Configuration (Intra-Frequency Handover)

Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off

How strong needed, for


RSRP Neighbour to
perform Intrafreq HO
Execution ?

Serving eNB Target eNB


Alamanda M-I-GF RC Veteran M-S-RT

www.piworks.net
Parameter Configuration (Intra-Frequency Handover)

Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off

www.piworks.net
Parameter Configuration (Intra-Frequency Handover)

Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off

www.piworks.net
Parameter Configuration (Intra-Frequency Handover)

Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off

www.piworks.net
Parameter Configuration (Intra-Frequency Handover)

Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off

Also called as CIO

www.piworks.net
Parameter Configuration (Intra-Frequency Handover)

Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off

www.piworks.net
Parameter Configuration (Intra-Frequency Handover)

Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off

How strong needed, for


RSRP Neighbour to
perform Intrafreq HO
Execution ?

Serving eNB Target eNB


Alamanda M-I-GF RC Veteran M-S-RT

Mn + 0 dB + 0 dB – 1 dB > -90 dBm + 0 dB + 0 dB + 1 dB


Mn > -90 dBm + 2 dB
Mn > -88 dBm

So, need more than +2 dB to perform Intrafreq HO in event A3

www.piworks.net
Parameter Configuration (Intra-Frequency Handover)

To reduce Ping Pong effect, there is


time to trigger to allowed Intrafreq HO

www.piworks.net
Connected Mode Intra Freq LTE – Turk Telekom

LTE/VoLTE/ViLTE 3 dB
L800 L800
Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off 3 dB
2 dB more strong L1800 L1800
2 dB
L2600 L2600
A B

With A3Off (2dB) and A3Hys (1dB). Handover in the cell edge will be halt
since Offset give UE more immunity power (2dB) than neighbor cell. With
reduction of frequent Handover on the cell edge will bring VoLTE service
Sub Parameter Default New
interruption less which can improve end user experience Parameter Name Remarks Target
Name Value Value
INTRAFREQHOA3HY 2 (0.5 2 (0.5
ST dBm) dBm)
All Local Cell ID for
INTRAFREQHOA3OF 2 (0.5 2 (0.5 No Change due to less site
L2600, Group ID:0 &
FSET dBm) dBm) amount
Group ID:1
INTRAFREQHOA3TI
320ms 320ms
METOTRIG
INTRAFREQHOA3HY 2 (0.5 2 (0.5
ST dBm) dBm)
All Local Cell ID for
MOD INTRAFREQHOA3OF 2 (0.5 4 (0.5 Immunity 3 dB from other
L1800, Group ID:0 &
Mn-Hys > Ms+Off INTRAFREQHOGROUP FSET dBm) dBm) cell
Group ID:1
-89 – (1) > -93 + (2) INTRAFREQHOA3TI
320ms 320ms
-90 > -91 METOTRIG
INTRAFREQHOA3HY 2 (0.5 2 (0.5
ST dBm) dBm)
INTRAFREQHOA3OF 2 (0.5 4 (0.5 Immunity 3 dB from other All Local Cell ID for L800,
FSET dBm) dBm) cell Group ID:0& Group ID:1
INTRAFREQHOA3TI
320ms 320ms
METOTRIG

www.piworks.net 29
Parameter Configuration (Inter-Frequency Handover)

• Inter-frequency handover triggering process Source Target


UE
Event A2 triggers the GAP measurement. eNB eNB
A3/A1/A2 Measurement
1.
Control
Two GAP modes
Measure RSRP/RSRQ
In a period of 40 ms (default) or 80 ms

Ms  Hys  Thresh A2 EventTriger

2. A2 Report

3 A4 Control

Event A4 triggers the inter-frequency handover.


GAP measurement

A4 EventTriger

4 A4 Reports

Handover preparation
Mn  Ofn  Ocn  Hys  Thresh

www.piworks.net
Connected Mode Inter Freq LTE & Inter RAT

LTE / ViLTE VoLTE


A1/A2/A4 A1/A2/A4
-110/-114/-110 -108/-112/-108
VoIP Priority
L1800 8 L1800
A1/A2/A4 A1/A2/A4 A1/A2/A4 A1/A2/A4 0
-108/-112/-108 -105/-109/-105 -108/-112/-108 -108/-112/-108
4
L800 L2600 L800 4 L2600
8
0

IRAT A1/A2 Fast Fast IRAT A1/A2 Cell Cell


IRAT A1/A2 IRAT A1/A2 IRAT A1/A2 IRAT A1/A2
-115/-126 Return Return -115/-126 Resel Resel
-115/-126 -108/-118 -108/-118 -108/-118
BLINDHO A1A2 BLINDHO A1A2 BLINDHO A1A2 UTRAN B1 UTRAN B1 UTRAN B1
-121 -121 -121 -103 -103 -103

U2100 U2100 U2100 U2100 U2100 U2100 8

SRVCC
Priority
U900 U900 U900 U900 U900 U900 2
PS REDIRECTION / CSFB SRVCC

www.piworks.net 31
User Distribution based on TT Layer Management

VoLTE User Steering feature


allowing eNB to assign L800 to
put more user during mobility

LTE PS User are didn’t


use steering. So more
user are camp on L1800

www.piworks.net
Mobility Performance

IntraFreq HOSR InterFreq HOSR


100% 100%

www.piworks.net
Signaling Flow for Handover

www.piworks.net
Mobility Performance (Measurement Report)

In this case there are Event A3


measurement report. Which
showing Neighbor RSRP(-72
dBm) bigger than Serving RSRP(-
77 dBm)

www.piworks.net
Mobility Performance (RRC Connection
Reconfiguration)

Important thing in this Layer 3


message there are: Target PCI,
Target DL EARFCN, RSI
Information,T304 and RS Power
from Target eNB

www.piworks.net
Mobility Performance (RRC Connection
Reconfiguration)

Important thing in this Layer 3


message there are: Preamble
Index assignment, transmission
mode, TDD Ratio configuration,

www.piworks.net
Mobility Performance (RRC Connection
Reconfiguration)

Important thing in this Layer 3


message there are: Chippering
and Integrity Protection Algorithm

www.piworks.net
Mobility Performance (RRC Connection
Reconfiguration Complete)

Important thing in this Layer 3


message there are: Information of
successful Handover Process.
And on below that message
contain RACH Contention Free
information during HO Process.

www.piworks.net
Conclusion of Mobility Performance
• This method should be taken from long call Logfile.
• 3 Layer message available for analysis.

Intrafreq

Interfreq

www.piworks.net
1.3 Drivetest Reporting on Integrity Performance

www.piworks.net
What is Integrity Performance ?

Why Internet in Indonesia so slow?

www.piworks.net
CA Layer Management – Turk Telekom

PCC / SCC Priority


Alpha L2600 4
SCell Blind

L1800 6
Beta SCell Blind SCell Blind

L800 5
SCC 2: Beta L2600

Gamma
PCC: Beta L1800
Default Current
Parameter Definition
value value

SCC 1: Gamma L800 Carrier Aggregation


CarrAggrA2ThdRsrp -109 -114
A2 RSRP threshold
Adaptive Mode was Activated
Then SCell can be served from different sector

www.piworks.net 43
What is Integrity Performance ?
From each Network
Element has different
capacity pipeline,
causing Integrity
performance so bad

Mbp
500

s
1 Gbps 1 Gbps

www.piworks.net
What is Integrity Performance ?
Optimization LTE not
only RF Parts, so we
have to cooperate to
fix it on entire Network
Element

Gbps
4
1 Gbps 1 Gbps

www.piworks.net
Introduction of LTE Radio Protocol Stack

• Two Planes in LTE Radio Protocol:  Main Functions of Control-plane:


o User-plane: For user data transfer  RLC and MAC layers perform the same functions
as for the user plane
o Control-plane: For system signaling transfer
 PDCP layer performs ciphering and integrity
protection
• Main Functions of User-plane:  RRC layer performs broadcast, paging, connection
management, RB control, mobility functions, UE
o Header Compression measurement reporting and control
o Ciphering  NAS layer performs EPS bearer management,
o Scheduling authentication, security control
Control-plane protocol stack
o ARQ/HARQ UE eNB MME

User-plane protocol stack NAS NAS

UE eNB RRC RRC

PDCP PDCP PDCP PDCP

RLC RLC RLC RLC

MAC MAC MAC MAC

PHY PHY PHY PHY

www.piworks.net
Physical Layer DL Throughput

www.piworks.net
Physical Layer DL Throughput

www.piworks.net
Physical Layer DL Throughput

www.piworks.net
Physical Layer DL Throughput

www.piworks.net
Throughput Chart

www.piworks.net
DL Throughput Performance

www.piworks.net
Transmission Mode

www.piworks.net
MIMO

MIMO is developed to provide doubled and more spectral efficiency. As an extension of singlei nput
single-output (SISO), MIMO uses multiple antennas at the transmitter and/or receiver in
combination with some signal processing techniques. Generally speaking, single-input multiple output
(SIMO), multiple-input single-output (MISO), and beamforming also belong to the
MIMO category.

www.piworks.net
MIMO Cont’d

www.piworks.net
Transmission Mode Configuration

OL Adaptive will be
focus on TM2 or TM3
only. Which no UE
feedback

www.piworks.net
Transmission Mode Configuration

No Adaptive will force


selection of TM.
OL (TM2/TM3)
CL (TM4/TM6)
OC
(TM2/TM3/TM4/TM6)

www.piworks.net
Transmission Mode Configuration

This MML Command will


force all area to be serve
with TM3 after Initial
Access

www.piworks.net
Close Loop Transmission Mode Trial

www.piworks.net
Close Loop Transmission Mode Trial

www.piworks.net
Close Loop Transmission Mode Trial

www.piworks.net
Close Loop Transmission Mode Trial

www.piworks.net
Beam Forming Configuration (TDD)

Beam Forming MIMO


are active by using
adaptive configuration

www.piworks.net
Beam Forming Configuration (TDD)

Beam Forming MIMO


are active by using
adaptive configuration

www.piworks.net
Beam Forming Configuration (TDD)

The setting of this parameter determines the transmission mode that applies to UEs served by
a multi-antenna eNodeB. The transmission modes vary with the application scenarios:
1. TM2 applies to UEs with low SINR and a high speed.
2. TM3 applies to UEs with high SINR and a high speed.
3. TM4 applies to UEs with high SINR and a low speed.
4. TM7 applies to UEs with low SINR and a low speed.
5. TM8 applies to UEs with high SINR and a low speed.
6. TM9w/oPMI applies to UEs with high SINR and a low speed.
7. TM9wPMI applies to UEs with high SINR and a low speed.
The throughput and coverage achieve the best performance only when the setting of this
parameter match the SINR and speed mobility of UEs. Otherwise, the performance is worse
than that with the optimal transmission mode.

www.piworks.net
Beam Forming Animation (TDD)

www.piworks.net
Mac DL Throughput Result

www.piworks.net
Low DL Throughput Issue due to Transmission
Normal Sites:
1. Hulu Gajah M-P-RT
2. Merbabu 7A M-P-RT
3. Rengas Permata M-
P-RT

Problematic Sites:
1. Jumhana M-S-
GF
2. Sutrisno M-T-RT

www.piworks.net
Low DL Throughput Issue due to Transmission

Problematic Sites:
1. Jumhana M-S-
GF
2. Sutrisno M-T-RT

www.piworks.net
Low DL Throughput Issue due to Transmission
Speed test on Sutrisno M-T-RT

www.piworks.net
Low DL Throughput Issue due to Transmission
FTP DOWNLOAD TEST UDP DOWNLOAD TEST

www.piworks.net
Summary & Recommendation

❖ Summary
- According to DT result, the SINR is very good. But DL throughput is very bad.
- Re-DT 2 times already (with different team), and the result is same. No issue
with DT Team & tools.
- Interference level on the bad throughput sites is normal.
- While performing FTP test (Mobility) & speed test (Static), the result is bad.
- While performing UDP test (Mobility), the result is good (Normal).
- No issue found on transmission part (Verified already with Bolt Team &
Drive Test result).

❖ Root Cause Suspect


- We assume there is protocol FTP, HTTP limitation on Datacom part which is
cannot distribute huge download throughput compare with UDP Test.
- There is some problem on switch/router setting regarding with DL/UL
Bandwidth

❖ Recommendation
- Need to escalate to Datacom/Core Team to check.

www.piworks.net
UDP Test

Log in to the LMT eNode-B

www.piworks.net
UDP Test

- Make sure UE be handled by right Local Cell ID


- To view these configuration typed “DSP CELL”

www.piworks.net
UDP Test

- Running the command “DSP ALLUEBASICINFO: LocalCellId=4;”


- The result is you can view that UE’s connected to the Locell ID
- Copy last 10 digit number

www.piworks.net
UDP Test

- TMSI Last 10 Digit Number : 3406059797


- IP Address UE : 172.25.24.35

www.piworks.net
UDP Test

- Running the command : “STR UUDATATST: UeIdType=STMSI_TYPE,


MmeCode=99, MTmsi=3406059797, ErabId=5, SrcIp="192.168.1.2",
SrcPort=6001, DstIp="172.25.24.35", DstPort=6001, PktSize=1400,
Rate=100000, Time=1200;”
- Upgrade the Rate to the 100000 (byte)
www.piworks.net
Low DL Throughput Issue due to UL Interference
PUCCH Flex Cfg Switch

• When the PucchFlexCfgSwitch option is selected, n RBs (n = PUCCHCFG.PucchExtendedRBNum)are moving from each of
the band edges to the middle. The original RBs are reserved for the PUSCH, and the PUCCH is allocated RBs from the (n +
1)th RB. Assume that the band consists of 100 RBs (numbered from 0 to 99) and the value of
PUCCHCFG.PucchExtendedRBNum is 5.Then, the PUCCH is allocated RBs from RB 5 and RB 94. After RBs are reserved
for the PUSCH, maximum RBs required for the PUCCH are allocated from both edges in the remaining RBs. Then, the PRACH
is allocated six continuous RBs from both edges in the remaining RBs,

• lists the maximum number of extended PUCCH RBs for each system bandwidth under the default physical resource
configurations. If this parameter is set to a value larger than the recommended value, the available system bandwidth cannot
meet the requirements of the preceding physical resource configurations and the cell cannot be set up or will be deactivated.

www.piworks.net 78
Low DL Throughput Issue due to UL Interference
Limitation on PUCCH Extended RB Number

If the PUCCH flexible configuration switch is turned on, interference to the PUCCH can
be reduced, which improves the PUCCH demodulation performance and also increases
the system access performance and DL throughput. However, the PUSCH resources
are separated, which decreases the UL throughput. Therefore, UL peak rate tests are
not supported in this scenario.
If this switch is turned off, two edges of the uplink frequency band may experience
severe interference, which decreases the PUCCH demodulation performance.

PucchExtendedRBNum:
Reserving more RBs reduces interference to the PUCCH, improves PUCCH
demodulation performance, enhances system access performance, and increases DL
throughput. However, it makes the PUSCH no longer continuous and decreases UL
throughput.
Configured based on the real RB interference situation.

www.piworks.net 79
PUCCH Structure

• Each half frame (6 OFDM Symbols) location of PUCCH are hopping from left edge to right edge from whole LTE Bandwidth. If
this RB are not clear from external interference it can be affected seriously to the Uplink ACK/NACK, CQI Reporting from UE to
eNB. So in this trial we have to find which RB are clear, so we decide to use shifting RB into 4 RB.

www.piworks.net 80
Summary Trial PUCCH Flex CFG SWC

• UL RSRP PUCCH Reduced 3 dB (Good Improvement)


• UL RSRP PUCCH Stable
• UL RSSI PUCCH Reduced 2 dB (Good Improvement)
• UL RSSI PUSCH Relative Increase 1 dB
• UL RSSI Increase contribute from PUSCH Area
• Improve Cell DL Throughput 10%

www.piworks.net 81
Low DL Throughput Issue due to UL Interference
UL RB Allocation 19 Dec 2016

PUCCH RB RIGHT(2RB next half frame)


PUCCH RB LEFT (2RB)

PRACH (6RB)

www.piworks.net
Low DL Throughput Issue due to UL Interference
UL RB Allocation 1 Jan 2017

Shifting 4RB from right to left


PUCCH RB RIGHT(2RB next half frame)
Shifting 4RB from left to right
PUCCH RB LEFT (2RB)

PRACH (6RB)
Shifting following PUCCH

www.piworks.net
Low DL Throughput Issue due to UL Interference

PUCCH RSSI reduced


About 6 dB

PUSCH RSSI increased


About 1 dB

www.piworks.net
Low DL Throughput Issue due to UL Interference

UL PUCCH RSRP reduced


About 1 dB

UL PUSCH RSRP reduced


About 1 dB

www.piworks.net
Low DL Throughput Issue due to UL Interference

UL Interference still exist on PUSCH

www.piworks.net
Low DL Throughput Issue due to UL Interference

Cell DL Throughput Improvement about 10% (1 Mbps)

Cell UL Throughput Relative stable

www.piworks.net
Low DL Throughput Issue due to UL Interference

User DL Throughput more stable on 8 Mbps

User UL Throughput Relative stable

www.piworks.net
DL PRB Utilization

High PRB Utilization for TD-LTE


15 MHz

www.piworks.net
RBG Allocation Strategy
Principle

PRB 49
RBG 16
PRB 48
PRB 47
RBG 15 PRB 46
PRB 45

PRB 2
RBG 1 PRB 1
PRB 0
1 Slot PRB 2
7 OFDMA Symbols RBG 0 PRB 1
0.5 ms PRB 0
*RBG System Bandwidth Of 10 MHz

www.piworks.net
RBG Allocation Strategy
Principle

Round Down (0) Round Down (0)


PRB 49 PRB 49 PRB 49 PRB 49
RBG 16 RBG 16 RBG 16 RBG 16
PRB 48 PRB 48 PRB 48 PRB 48
PRB 47 PRB 47 PRB 47 PRB 47
RBG 15 PRB 46 RBG 15 PRB 46 RBG 15 PRB 46 RBG 15 PRB 46
PRB 45 PRB 45 PRB 45 PRB 45
N(RBG)<1 N(RBG)+1
PRB 49 PRB 2 PRB 2 PRB 49 PRB 2 PRB 2
RBG 16 RBG 16
PRB 48 RBG 1 PRB 1 RBG 1 PRB 1 PRB 48 RBG 1 PRB 1 RBG 1 PRB 1
PRB 47 PRB 0 PRB 0 PRB 47 PRB 0 PRB 0
RBG 15 PRB 46 PRB 2 PRB 2 RBG 15 PRB 46 PRB 2 PRB 2
PRB 45 RBG 0 PRB 1 RBG 0 PRB 1 PRB 45 RBG 0 PRB 1 RBG 0 PRB 1
PRB 0 PRB 0 PRB 0 PRB 0

PRB 2
Round Up (1) PRB 2
Round Up (1)
RBG 1 PRB 1 RBG 1 PRB 1
PRB 49 PRB 49 PRB 49 PRB 49
PRB 0 RBG 16 RBG 16 PRB 0 RBG 16 RBG 16
PRB 48 PRB 48 PRB 48 PRB 48
PRB 2 PRB 2
PRB 47 PRB 47 PRB 47 PRB 47
RBG 0 PRB 1 RBG 0 PRB 1
RBG 15 PRB 46 RBG 15 PRB 46 RBG 15 PRB 46 RBG 15 PRB 46
PRB 0 PRB 0
PRB 45 PRB 45 PRB 45 PRB 45

PRB 2 PRB 2 PRB 2 PRB 2


RBG 1 PRB 1 RBG 1 PRB 1 RBG 1 PRB 1 RBG 1 PRB 1
PRB 0 PRB 0 PRB 0 PRB 0
PRB 2 PRB 2 PRB 2 PRB 2
RBG 0 PRB 1 RBG 0 PRB 1 RBG 0 PRB 1 RBG 0 PRB 1
PRB 0 PRB 0 PRB 0 PRB 0

www.piworks.net
LST CELLDLSCHALGO

www.piworks.net
Speed Test Before (RBG Round Down)

HUAWEI Mate 7 Galaxy J5

www.piworks.net 93
Speed Test After (RBG Round UP)

HUAWEI Mate 7 Galaxy J5 (Cat4)


Before After Rating Before After Rating

Ping Ping Ping Ping


30ms 19ms 19ms 18ms

DL DL DL DL
34.23 42.86 62.63 64.79

UL UL UL UL
19.84 20.64 42.11 41.46

DL Throughput has 25% Improvement DL Throughput has 4% Improvement


UL Throughput has no effect UL Throughput has no impact
Ping Test has 36% reduced latency Ping Test has 5% reduced latency

All related tests have been performed


in the TT Office (UMRANIYE).

Ping latency average & DL Throughput


have been improved by using DL SCH
RBG Round Up

www.piworks.net 94
Monitoring KPI

Round Up

Round Down

• Cell DL Avg Throughput KPIs have been improved around %10 after allocation strategy changed from ROUND_DOWN to
ROUND_UP.
• Insipite of the high data speed improvement, PRB utilization has not been affected by the change as much as expected.
Around 1 point increament has been experienced at the receipe cutoff date and the trend has came back to the previous
percentages afterward.

www.piworks.net 95
Monitoring KPI

Round Down

Round Up

• User TP is calculated by total bits except last TTI


bits over Uplink data transmit duration except the
last TTI before the UE buffer is empty.

User TP = 'L.Thrp.bits.DL‘ - 'L.Thrp.bits.DL.LastTTI') / ('L.Thrp.Time.DL.RmvLastTTI)

• KPI was created to simulate user behaviour in


order to discard the last small data at last TTI
period. Excluded bits have increased around 3
times with new algorithm although the duration
was same as before. Therefore, the formula can
not reflect the real user data speed because the
last bits are not least as much as before.

PRB Usage Rate should be monitored, when daily result of PRB Usage are
more than 80% this RBG Round Up should be roll back to Round Down

www.piworks.net 96
Summary Trial Smartpreallocation (Low UL Throughput)

DrxAlgSwitch = OFF DrxAlgSwitch = ON


KPI’s
SmartPreAllocation SmartPreAllocation SmartPreAllocation SmartPreAllocation
Duration = 0 ms Duration = 50 ms Duration = 0 ms Duration = 50 ms
Maintain (No
Cell UL Throughput Increased (Good) Decreased (Bad) Decreased (Bad)
Effect)
Maintain (No
UL PRB Used Rate Decreased (Good) Increased (Bad) Increased (Bad)
Effect)
RRC Setup Maintain (No
Increased (Good) Decreased (Bad) Maintain
Success Rate Effect)
ERAB Setup Decreased Maintain (No Increased (Good)
Increased (Good)
Success Rate (Maintain Level) Effect)
Maintain (No
UL Interference Decreased (Good) Increased (Bad) Increased (Bad)
Effect)

www.piworks.net 97
(SmartPreAllocationDuration = 0ms) – DRX OFF

Summary:
• UL PRB Used Rate are
decreased (became less
occupied)
• Cell UL Throughput are
increased (UE less
Interference and have
more vacant UL PRB,
addition 25% according
parameter
PreAllocationBandwidthR
atio)

www.piworks.net 98
(SmartPreAllocationDuration = 50ms) – DRX OFF

Summary:
• UL PRB Used
Rate are
increased
(became congest)
• Cell UL
Throughput are
decreased (UE
increase the
power transmit
and occupied
more UL resource)

www.piworks.net 99
(SmartPreAllocationDuration = 0ms) – DRX ON

Summary:
• UL PRB Used Rate are
maintain (no effect)
• Cell UL Throughput are
maintain (no effect)

www.piworks.net 100
(SmartPreAllocationDuration = 50ms) – DRX ON

Smartpre: ON

Summary:
• Cell Downlink Avg. Throughput shows there
is no much significant improvement (related
with the RF Condition (Overshoot Issue) and
User Behavior (only speedtest required huge
throughput).
Smartpre: ON
• Cell Uplink Avg. Throughput show there is
degradation on avg. lower value of UL
Throughput (same with speedtest result), on
the other hand there is huge traffic volume
on UL side which increase the Avg.Uplink
Throughput on several hour.

www.piworks.net 101
LST DRX

www.piworks.net
LST CELLALGOSWITCH

www.piworks.net
DL Scheduling Parameter

Can be set as Using DL MIMO


EPF/PF/RR/MaxCI 2x2

www.piworks.net
Antenna Configuration

Antenna set as
4T4R

www.piworks.net
Rank Indicator

In simple words, RI is an indicator showing how well multiple Antenna work. What do you mean by "how well the multiple Antenna work ?". We
usually say "Each of the multiple antenna (e.g, each antenna in MIMO configuration) works well if the signal from each antenna has NO correlation
to each other". "No correlation" implies "no interference to each other".
Maxmum RI value is very closely related to the number of Antenna. Maximum RI is same as number of antenna on each side if the number of Tx
antenna and Rx antenna is same. If the number of Tx and Rx are different, the one with less antenna is the same as Max achievable RI.

Max RI means "No Correlation between the antenna", "No interference to each other", "Best Performance".
For example, in case of 2x2 MIMO, the RI value can be 1 or 2. When the value 2 in this case means "No Correlation between the antenna", "No
interference to each other", "Best Performance". If the value is 1, it implies that the signal from the two Tx antenna is percieved by UE to be like
single signal from single Antenna, which means the worst performance.

www.piworks.net
DL MIMO
There’s an additional benefit as well. A Gigabit LTE device has four antennas in order to support 4x4
MIMO, giving it a hidden edge. In good signal conditions, you can get four streams of data that increase
your speed, as compared to two streams with conventional LTE. In weak signal conditions, the additional
antennas act like additional “ears” that are designed to help your Gigabit LTE device lock on to the signal
from the tower, which can yield up to 70 percent faster speeds.
Modem Name
•Qualcomm® Snapdragon™ X16 LTE modem
LTE Category
•LTE Category 16 (downlink)
•LTE Category 13 (uplink)

Downlink Features
•4x20 MHz carrier aggregation
•Up to 256-QAM
•DL MIMO 4x4

Uplink Features
•Qualcomm® Snapdragon™ Upload+
4T4 •2x20 MHz carrier aggregation
R •Up to 64-QAM

Peak Download Speed


•1 Gbps
Peak Upload Speed
•150 Mbps

Supported Cellular Technologies


•LTE FDD
•LTE TDD
•LTE-U
•LAA
•LTE Broadcast
Qualcomm MSM8998 Snapdragon 835 •WCDMA (DB-DC-HSDPA, DC-HSUPA)
•TD-SCDMA
•CDMA 1x
•EV-DO
•GSM/EDGE
4x4 Multi SIM
•LTE Dual SIM Dual Standby (DSDS)
Next-generation Calling Services
•VoLTE with SRVCC to 3G and 2G
•HD and Ultra HD Voice (EVS)
•CSFB to 3G and 2G

www.piworks.net 107
LST CELL

www.piworks.net
LST CELLDLSCHALGO

www.piworks.net
LST ENODEBALGOSWITCH

In downlink 4x4 MIMO, a


maximum of four layers can be
scheduled. If UE categories are 6
or higher (excluding 8 and 14),
the scheduling ranks are
controlled by the
Tm3Tm4Max4LayerCtrlSwitch
option of the
eNodeBAlgoSwitch.Compatibil
ityCtrlSwitch parameter:

•If this option is selected, the


scheduling ranks can be ranks 1
to 4. (Proposed to be ON)

•If this option is deselected, the


scheduling ranks can only be
rank 1 and 2.

Tm3Tm4Max4LayerCtrlSwitch: Indicates whether UEs of CAT6 and above whose UE capabilities exclude CAT5 support 4-layer MIMO based on TM3/TM4. If this option is selected, UEs of CAT6 and above whose UE capabilities
exclude CAT5 support MIMO at a maximum of four layers based on TM3/TM4. If this option is deselected, UEs of CAT6 and above whose UE capabilities exclude CAT5 support MIMO at a maximum of two layers based on
TM3/TM4.

www.piworks.net
UE LTE Category (Tech Spec 36.306 v14.1.0)

Downlink Uplink
Maximum number of DL- Maximum Maximum number of UL-SCH
Maximum
SCH transport block bits number of transport block bits transmitted
number of bits
3GPP received within a TTI Maximum number of bits Total number of supported Support for within a TTI
Category of an UL-SCH Support for
release of a DL-SCH transport soft channel layers for 256QAM in
transport block 64QAM in UL
block received within a TTI bits spatial DL
(Mbit/s) (Mbit/s) transmitted
multiplexing in
within a TTI
DL
0 Rel 12 1000 1 1000 25344 1 No 1000 1 1000 No
1 Rel 8 10296 10 10296 250368 1 No 5160 5 5160 No
1bis Rel 14 10296 10 10296 250368 1 No 5160 5 5160 No
2 Rel 8 51024 51 51024 1237248 2 No 25456 25 25456 No
3 Rel 8 102048 102 75376 1237248 2 No 51024 51 51024 No
4 Rel 8 150752 150 75376 1827072 2 No 51024 51 51024 No
5 Rel 8 299552 299 149776 3667200 4 No 75376 75 75376 Yes
75376 (2 layers, 64QAM)
6 4 Rel 10 301504 301 3654144 2 or 4 No 51024 51 51024 No
149776 (4 layers, 64QAM)
75376 (2 layers, 64QAM)
7 4 Rel 10 301504 301 3654144 2 or 4 No 102048 102 51024 No
149776 (4 layers, 64QAM)
8 5 Rel 10 2998560 2998 299856 35982720 8 No 1497760 1497 149776 Yes
75376 (2 layers, 64QAM)
9 6,4 Rel 11 452256 452 5481216 2 or 4 No 51024 51 51024 No
149776 (4 layers, 64QAM)
75376 (2 layers, 64QAM)
10 7,4 Rel 11 452256 452 5481216 2 or 4 No 102048 102 51024 No
149776 (4 layers, 64QAM)
A UE indicating category 6 or 7 shall also indicate category 4
A UE indicating category 8 shall also indicate category 5
A UE indicating category 9 shall also indicate category 6 and 4
A UE indicating category 10 shall also indicate category 7 and 4
A UE indicating category 11 shall also indicate category 9, 6 and 4
A UE indicating category 12 shall also indicate category 10, 7 and 4

www.piworks.net 111
Downlink
Maximum number of DL-SCH transport block bits received within a TTI Maximum number of
Downlink 3GPP Maximum number of bits of
Total number of supported layers for Support for
category release a DL-SCH transport block
soft channel bits spatial multiplexing in 256QAM in DL
(Mbit/s) received within a TTI
DL
M1 Rel 13 1000 1 1000 25344 1 No
M2 Rel 14 4008 4 1000 73152 1 No
0 Rel 12 1000 1 1000 25344 1 No
1bis Rel 14 10296 10 10296 250368 1 No
75376 (2 layers, 64QAM)
6 Rel 12 301504 301 3654144 2 or 4 No
149776 (4 layers, 64QAM)
75376 (2 layers, 64QAM)
7 Rel 12 301504 301 3654144 2 or 4 No
149776 (4 layers, 64QAM)
75376 (2 layers, 64QAM)
9 Rel 12 452256 452 5481216 2 or 4 No
149776 (4 layers, 64QAM)
75376 (2 layers, 64QAM)
10 Rel 12 452256 452 5481216 2 or 4 No
149776 (4 layers, 64QAM)
75376 (2 layers, 64QAM)
97896 (2 layers, 256QAM)
11 Rel 12 603008 603 7308288 2 or 4 Optional
149776 (4 layers, 64QAM)
195816 (4 layers, 256QAM)
75376 (2 layers, 64QAM)
97896 (2 layers, 256QAM)
12 Rel 12 603008 603 7308288 2 or 4 Optional
149776 (4 layers, 64QAM)
195816 (4 layers, 256QAM)
97896 (2 layers, 256QAM)
13 Rel 12 391632 391 3654144 2 or 4 Mandatory
195816 (4 layers, 256QAM)
14 Rel 12 3916560 3916 391656 (8 layers, 256QAM) 47431680 8 Mandatory
75376 (2 layers, 64QAM)
97896 (2 layers, 256QAM)
15 Rel 12 749856 - 798800 749 - 798 9744384 2 or 4 Optional
149776 (4 layers, 64QAM)
195816 (4 layers, 256QAM)
75376 (2 layers, 64QAM)
97896 (2 layers, 256QAM)
16 Rel 12 978960 - 1051360 978 - 1051 12789504 2 or 4 Optional
149776 (4 layers, 64QAM)
195816 (4 layers, 256QAM)
17 Rel 13 25065984 25065 391656 (8 layers, 256QAM) 303562752 8 Mandatory
75376 (2 layers, 64QAM)
97896 (2 layers, 256QAM)
149776 (4 layers, 64QAM)
18 Rel 13 1174752 - 1206016 1174 - 1206 14616576 2 or 4 or 8 Optional
195816 (4 layers, 256QAM)
299856 (8 layers, 64QAM)
391856 (8 layers, 256QAM)
75376 (2 layers, 64QAM)
97896 (2 layers, 256QAM)
149776 (4 layers, 64QAM)
19 Rel 13 1566336 - 1658272 1566 - 1658 19488768 2 or 4 or 8 Optional
195816 (4 layers, 256QAM)
299856 (8 layers, 64QAM)
391856 (8 layers, 256QAM)

www.piworks.net 112
DL MIMO 4x4 Trial – Turk Telekom

www.piworks.net
DL MIMO 4x4 Trial – Turk Telekom

www.piworks.net
DL MIMO 4x4 Trial – Turk Telekom

www.piworks.net
DL MIMO 4x4 Trial – Turk Telekom

www.piworks.net
CQI Comparison

Afternoon 16:00 Midnight 02:00

www.piworks.net
CQI Comparison

www.piworks.net
DL Throughput Comparison
Afternoon 16:00 Midnight 02:00

During Low Utilization of DL PRB on Cell Level give benefit for single UE
can get higher DL PRB which is related to the Higher DL Throughput.
www.piworks.net
DL Throughput Comparison

During Low Utilization of DL PRB on Cell Level give benefit for single UE
can get higher DL PRB which is related to the Higher DL Throughput.
www.piworks.net
CQI to MCS

CQI stands for Channel Quality Indicator. As the name implies,


it is an indicator carrying the information on how good/bad the
communication channel quality is. This CQI is for HSDPA. (LTE
also has CQI for its own purpose).

CQI is the information that UE sends to the network and


practically it implies the following two
i) Current Communication Channel Quality is this-and-that..
ii) I (UE) wants to get the data with this-and-that transport block
size, which in turn can be directly converted into throughput

www.piworks.net
Conclusion of Integrity Performance
• This method should be taken from long call Logfile.
• There is no specific Layer 3 message available for analysis.

DL Throughput
UL Throughput

www.piworks.net
1.4 Drivetest Reporting on Retainability Performance

www.piworks.net
Measurement Event

Drop event happen after UE


have been connected.
RRC Attempt
RRC Connected

Drop Event

www.piworks.net
Drop Rate Calculation

Drop Rate % :
RRCConnection_ReestablishmentFailed/RRCConnectionSe
tupOK*100%

Drop Rate : 1/116*100%


: 0.86%

www.piworks.net
RRC Reestablishment Request

www.piworks.net
RRC Reestablishment Reject

From the drive test result shown that, minimum RSRQ value are below the
allowable minimum threshold at -18 dB (U2000 Configuration)

www.piworks.net
Cell Selection Configuration

www.piworks.net
Idle Mode Turk Telekom

L2L Low

L2L High
L2W Low

www.piworks.net 129
Conclusion of Retainability Performance
• This method should be taken from short call Logfile.
• 3 Layer message available for analysis.

www.piworks.net
End of Section
Hoşçakalın !

www.piworks.net
Author

Ray Khastur, ST.


Educational Background :
▪ Bachelor of Telecommunication Engineer, Telkom University – INDONESIA
(2006 – 2010) JPPA-N Acceleration

Professional Experience :
▪ PI.Works, Senior RNO Consultant (2017 ~ Present)
▪ PI.Works, RAN Consultant (2014 ~ 2017)
▪ China JIESAI, LTE Optimization Consultant (2014)
▪ Lintas Media Telekomunikasi, LTE RF Team Leader (2013 ~ 2014)
▪ HUAWEI Tech Investment, CWiL RNP/O Engineer (2011 ~ 2013)
▪ Transdata Global Network, CDMA RNO Engineer (2010 ~2011)

Achievements :
▪ Turk Telekom HUAWEI LTE Network Improvement Consultant – 2016 PIW
▪ RF Network Design & Planning First Commercial LTE Network in Indonesia
(BOLT! Super 4G LTE)-2013 HUAWEI
▪ NPI & Post Launch Optimization BOLT! Super 4G LTE - 2014 PI.Works
▪ LTE Technical Expert for HUAWEI Environment
▪ LTE Planning & Optimization Trainer Floatway Certified-2013
▪ LTE HUAWEI Trainer for Subcont-2016

www.piworks.net

You might also like