Ericsson Confidential

GUIDE TO CHECK SETTINGS/I


12 (12)

Prepared (also subject responsible if other)

No.

EEM/GMW ECECREI

Approved

Checked

Date

Rev

Reference

2009-02-16

PA2








GUIDE TO CHECK SETTINGS/ISSUES WITH HSDPA TP


1 Introduction 2

2 Configuration 2

2.1 Perform an HSDPA Parameter audit, and make sure that all HS parameters are according to. 2

2.2 SGSN issues 3

2.3 RBS number of codes is enough 3

2.4 IP router issues 4

3 Hardware 4

4 KPIs and counters 4

5 Transmission 8

5.1 Limit on Bandwidth 8

5.2 Different APNs 8

5.3 Unestable transmission 8

6 Radio conditions 9

6.1 Bad RF conditions. 9

6.2 Uplink RSSI 9

7 Traces 9

7.1 Analyse Layer3 messages for dropcall, setup failure, PDP context reject, and make sure they are not contributing to the low throughput. 9

7.2 MBR in Rab-Assignment that UE received. 9

8 UE issues 10

9 Channel Elements 12



1Introduction

Throughput is used to describe the data rate. It is possible to measure throughput for each protocol layer separately, such as MAC-hs and RLC layers as well as for TCP/IP. Each protocol layer adds header and functionality affecting performance, why it is important to understand the implications and limitations for each measurement. Below is a table to visualize the differences between UE categories and throughput limitations per protocol layer.



Table 1 Maximum throughput per protocol layer.

Maximum provided (Mbps):

Category 12

Category 6

Category 8

Category 10

#Codes

5

5

10

15

Modulation

QPSK

16QAM

16QAM

16QAM

UE Supported HS-DSCH rate

1.82

3.65

7.21

14.0

MAC-hs layer (incl header)

1.72

3.58

6.95

13.8

MAC-hs layer (excl. MAC-hs header, padding, 10% MAC-hs BLER)

1.51

3.18

6.20

12.4

Theoretical maximum RLC user data throughput

1.60

3.36

6.72

13.4

RLC layer (excl. RLC header, 1% RLC signaling)

1.43

2.99

6.00

12.0

Application layer
(excl. TCP/IP header)

1.39

2.91

5.83

11.7

2Configuration

 

2.1Perform an HSDPA Parameter audit, and make sure that all HS parameters are according to.

a- The “RNC Managed Object Model” (190 89-AXD 105 03/1 Uen AE4A) provided in the CPI unless these parameters are modified for optimization purposes.:

b- Transport Network Configuration (1/1553-HSD 101 02/5 Uen H).

c- Radio Network Guideline (100/1553-HSD 101 02/5 Uen A)

d- HSDPA User Plane (93/1553-HSD 101 02/5 Uen A)

e- HSDPA Migration and activation (95/1553-HSD 101 02/5 Uen B)

e- Special RNC parameters: cqiFeedbackCycle, deltaAck1, deltaAck2, deltaCqi1, deltaCqi2, deltaNack1, deltaNack2 

moshell> get Hsdsch cqif | delta

f- Special RBS parameters: cqiErrors, cqiErrorsAbsent, hsPowerMargin, hsScchMaxCodePower, queueSelectAlgorithm| maxHsRate

moshell> get carr cqi | scch | queu

2.2SGSN issues

Check that ATM  PCR and SCR value are set according to values in MGS and RNC. If values are too low in IuU interface SGSN and it does not match PCR value in MGW and  RNC, this will directly affect HS TP.

Also check that pdr devices are enabled in RNC and SGSN sides.

Moshell> st pdr

2.3RBS number of codes is enough

Check that the number of codes needed for the UE are available.

Except the MaxHsRate in RBS, Category 7 - 10 UE requires more than 5 codes in one cell, so the DynamicCodeAllocation should be activated, or manually allocate more codes in the cell.

Dynamic Code Allocation automatically adapts the HSDPA code allocation to the DCH traffic load, favoring e.g. voice traffic in busy hour. For RBS base band hardware configurations where 15 HSDPA codes are shared in several sectors this feature also adapts to asymmetric sector capacity, i.e. by automatically allocating more HSDPA codes to a sector with more HSDPA users.

Parameters: There will be operator configurable parameters on cell level to control the minimum HSDPA code allocation:

> set RbsLocalCell dynamicHsPdschCodeAdditionOn true

> set utrancell=xxx numHsPdschCodes 1

2.4IP router issues

IP router settings can affect the TP in RBSs. E.g. Changes in load balancing settings.

3Hardware

Check HS board status with moshell command cabx:

rbs> cabx

Also check te log read on board for errors.

Rbs> te log read

Rbs> lhsh 00XX00 te log read

Check In RAXB, constant pathfinder error message.

Rbs> lhsh 00XX00 te log read

RAX, ReadErrorLog: DEC_PathFinderIntr swc/io/src/pf.c:255 INFO:ABN: Pathfinder error. Status reg hi=0x8000000 low=0xa8000 : 0x247E

Replace RAX board.

4KPIs and counters

Collect the following counters to get an overview of the HSDPA/throughput performance on the network

RNC Counters:

Counters for HSDPA Drops:

pmNoSystemRbReleaseHs - HSDPA RAB drops (system or abnormal releases) per cell. Drop rate is calculated as 'Dropped RABs/Successfully established RABs'.
pmNoNormalRbReleaseHs - HSDPA RAB normal disconnections per cell.
pmInactivityHsIdle - The number of signaling connection releases triggered for PS Interactive RAB mapped on HS-DSCH due to inactivity.

Counters for HSDPA RAB establishments: 

pmNoRabEstablishAttemptPacketInteractiveHs - HSDPA RAB establishment attempts per cell.
pmNoRabEstablishSuccessPacketInteractiveHs - HSDPA RAB successful establishments per cell.

Counters for HSDPA throughput:

pmCapAlloclubHsLimitingRatio – Percentage of time when capacity allocation was limited by the Iub bandwidth.
pmHsDataFramesReceived – total number of HS frames received over Iub interface.
pmIubMacdPduRbsReceivedBits – Received numbers of Iub MAC-d PDU bits per sec.
pmTargetHsRate – Target HS rate as percentage of the value of the maxHsRate parameter.
pmAverageUserRate – Distribution of average user rate among all users in HS-DSCH in a cell.

Counters for HSDPA handovers: 

pmNoHsCcAttempt - Number of attempted Serving HS-DSCH Cell Change counted on target cell.
pmNoHsCcSuccess - Number of successful Serving HS-DSCH Cell Change counted on target cell.

Counters for HSDPA traffic levels: 

pmSumPsHsAdchRabEstablish/pmSamplesPsHsAdchRabEstablish = Average number of A-DCH connections per cell. Scans are done every 5 seconds. pmSumBestPsHsAdchRabEstablish/pmSamplesBestPsHsAdchRabEstablish = Average number of A-DCH connections counted per best cell. Scans are done every 5 seconds.
pmTransmittedAtmCells – Number of transmitted ATM cells through the ATM port. The counter is available for CPP (RNC and RBS).
PmReceivedAtmCells - Number of received ATM cells through the ATM port. The counter is available for CPP (RNC and RBS).
pmSecondsWithUnexp – Errored seconds due to protocol error. The counter is available for CPP (RNC and RBS).

RNC Moshell :

Moshell command pmr gives the throughput on the RNC level, however these values can't be used to troubleshoot specific sites:
Refer to the RBS counter to get the throughput on cell level.

Rnc_rbs> pmr

Rnc_rbs> pget mo counter

Rnc_rbs> pmx mo counter

The difference between HsTputAverg and HsTputGt100KbAverg is as follows:

HsTputAverg = 8 * 1000 * ( pmSentPacketDataHs1 / 1000 + pmSentPacketDataHs2 / 1000 + pmSentPacketDataHs3 + pmSentPacketDataHs4 ) / ( pmTotalPacketDurationHs1 + pmTotalPacketDurationHs2 + pmTotalPacketDurationHs3 + pmTotalPacketDurationHs4 )

HsTputGt100KbAverg = 8 * 1000 * (pmSentPacketDataHs4) / (pmTotalPacketDurationHs4)

where:

pmSentPacketDataHs1 is the total amount of sent data excluding retransmissions (this counter is stepped only if the amount of data transmitted is < 1kB)
pmSentPacketDataHs2 is the total amount of sent data excluding retransmissions (this counter is stepped only if the amount of data transmitted is < 10kB but > 1kB)
pmSentPacketDataHs3 is the total amount of sent data excluding retransmissions (this counter is stepped only if the amount of data transmitted is < 100kB but > 10kB)
pmSentPacketDataHs4 is the total amount of sent data excluding retransmissions (this counter is stepped only if the amount of data transmitted is > 100kB).
pmTotalPacketDurationHs1 - is the duration for data < 1kB.
pmTotalPacketDurationHs2 - is the duration for data < 10kB but > 1kB

pmTotalPacketDurationHs3 - is the duration for data < 100kB but > 10kB
pmTotalPacketDurationHs4 - is the duration for data > 100kB

HsTputAverg is looking at HS throughput over all data transmission sizes, even if a small amount of data is sent, where as HsTputGt100KbAverg is for data transmissions of > 100kB. The HsTputGt100KbAverg should give a better indication of the performance achieved with HS - because you'd be most interested in the speed when performing large downloads rather than when using smaller transmission sizes where you may not be utilizing the full 2ms air interface frame.

RBS Counters and formulas:

pmSumAckedBits : The number of bits transmitted at MAC-hs and acknowledged by the UE
pmSumNonEmptyUserBuffers : The number of user buffers with data in the buffer for each 2 ms subframe
pmNoActiveSubFrame : The sum of active 2 ms subframes
pmSumTransmittedBits: The number of transmitted bits at MAC-hs level including retransmissions

- The average MAC-hs throughput per PS Interactive HSDPA user:

pmSumAckedBits / (TTI * pmSumNonEmptyUserBuffers)

- The MAC-hs throughput on cell level:

pmSumAckedBits / (TTI * pmNoActiveSubFrame)

- The MAC-hs data rate on cell level:

pmSumTransmittedBits / (TTI * pmNoActiveSubFrame)

where TTI = 0.002 s.

The following counters can also help when troubleshooting problems with low HSDPA thoughtput:

pmNoFailedRabEstAttemptLackDlAse. Number of failed RAB establishment attempts due to lack of DL ASE.

pmNoFailedRabEstAttemptLackUlAse. Number of failed RAB establishment attempts due to lack of UL ASE.


pmNoFailedRabEstAttemptLackDlChnlCode. Number of failed RAB establishment attempts due to lack of DL

pmNoFailedRabEstAttemptLackDlPwr. Number of failed RAB establishment attempts due to lack of DL power.

pmNoFailedRabEstAttemptLackUlAse. Number of failed RAB establishment attempts due to lack of UL ASE

5Transmission

5.1Limit on Bandwidth

In order to have a good throughput for HSDPA a key issue is the transmission design.

A limitation in the transmission (e.g. E1s in the Iub interface, STM-1 in Iu interface) will directly affect the maximum throughput available. Congestion may occur due to limited number of E1s.

In order to check if the Iub is congested the following counter can be checked:

pmCapAlloclubHsLimitingRatio – Capacity Allocation Iub HS Limiting Ratio.

This is the relative number of occurrences when the calculated Capacity Allocation figure is limited by the Iub HS bandwidth during a 100 ms period (compared to the total number of 100 ms periods in the PM interval).

5.2Different APNs

Sometimes when choosing different APN (Access point Name) in the UE, the throughput is different. E.g. If there is any transmission issue on one of connections for the Gn interface, a longer delay might be experienced by UE’s trying to connect through this path to the internet and this will provoke a lower throughput.

5.3Unestable transmission

Check errors in GPB. If errors like the following are seen:

ReadErrorLog: GPB , Cs_a50Server_proc a501_server.c:3543 INFO:Lost 100 packets on channel 326 due to Errors: Length CRC CPI CPUU

It could mean transmission problems.

Perform VclTp end to end loopback to check if this is stable or not.



6Radio conditions

6.1Bad RF conditions.

Optimization to be carried out in areas

6.2Uplink RSSI

Check stats in RBS.

Rbs> pmr

7Traces

7.1Analyse Layer3 messages for dropcall, setup failure, PDP context reject, and make sure they are not contributing to the low throughput.

Analyze if drops are related to specific IMSIs.

In some occasion SW faults have been found in phone vendors that have provoked KPI and TP degradations.

7.2MBR in Rab-Assignment that UE received.

The RAB-Assignment Message:

rAB-Parameters:

{ trafficClass interactive,
rAB-AsymmetryIndicator asymmetric-bidirectional,
maxBitrate
{ 2048000, >>> DL
384000 }, >>> UL


This assigned MBR is the consequence of negotiation between UE and several nodes – RBS, HLR, SGSN and GGSN…… So if this rate is LOWER than expected, verify QoS that defined in each node (Including UE):


8UE issues

Check that UE category supports HSDPA TP required.

























HS-DSCH category

Max bit rate with QPSK [Mbps]

Max bit rate with 16QAM [Mbps]

Category 1

0.69

1.12

Category 2

0.69

1.12

Category 3

1.04

1.68

Category 4

1.04

1.68

Category 5

2.08

3.36

Category 6

2.08

3.36

Category 7

4.32/4.48

6.72/6.72

Category 8

4.32/4.48

6.72/6.72

Category 9

6.56/6.72

9.6/9.6

Category 10

6.56/6.72

11.2/13.44

Category 11

0.8

-

Category 12

1.6

-

UE category can be checked in the following way:

9Channel Elements

Following counters in P5 can be used to check if RAB establishment has failed due to lack of resources.

Rbs> pmx utrancell pmNoFailedRabEstAttemptLack

pmNoFailedRabEstAttemptLackDlHw. Number of failed RAB establishment attempts due to lack of DL hardware resources

pmNoFailedRabEstAttemptLackDlHwBest Number of failed RAB establishment attempts due to lack of DL hardware resources, for the best cell in the active set.

pmNoFailedRabEstAttemptLackUlHw. Number of failed RAB establishment attempts due to lack of UL hardware resources

pmNoFailedRabEstAttemptLackUlHwBest. Number of failed RAB establishment attempts due to lack of UL hardware resources, for the best cell in the active set.

Some examples of known issues: