Access Failures Troubleshooting Workshop

Security Level:

Huawei Workshop Troubleshooting Access Failures

May 17th, 2011 www.huawei.com

HUAWEI TECHNOLOGIES CO., LTD.

Huawei Confidential

Contents •

Call Setup Procedure (step by step & all protocols)

•

General Causes of failures

•

How to chase and to solve specific access failures:  RRC Access Failure Troubleshooting.  Paging Access Failure Troubleshooting  RACH Access Failure Troubleshooting


Huawei Confidential

Page 2

Mobile Terminated Call Setup Procedure (I) UE

Node B

RNC

Here the UE will start to send the PRACH and wait for AICH and then send RACH message

MSC / VLR

ISUP

MGW 1. IAM

2. PAGING RANAP

RANAP

3. PCH: PCCH: PAGING TYPE 1 RRC

RRC 4. RACH: CCCH: RRC CONNECTION REQUEST

RRC

RRC

NBAP

5. RADIO LINK SETUP REQUEST

NBAP

Start RX

NBAP

ALCAP

Here the Node-B will start RL with DL transmission

ALCAP

DCH-FP

Can be either RRC Connection setup (to this cell and or inter freq to another one when DRD) or Reject.

6. RADIO LINK SETUP RESPONSE

7. ESTABLISHMENT REQUEST (AAL2)

8. ESTABLISHMENT CONFIRM (AAL2)

9. DOWNLINK SYNCHRONISATION

10. UPLINK SYNCHRONISATION

DCH-FP

Here the RNC will perform a DRD decision and CAC decision for RRC

NBAP

ALCAP

ALCAP

DCH-FP

DCH-FP

Start TX

11. FACH: CCCH: RRC CONNECTION SETUP RRC L1

RRC 12. SYNCH IND

Here the UE will do DL synchronization (using N312=1, T312=1, N313=20 andT313=3) . Then the UE will start UL TX transmission and the Node-B will detect UL SYNCH (based on N_INSYNCIND=8, N_OUTOFSYNCIND=8,TRLFAILURE=20)

L1

NBAP

13. RADIO LINK RESTORE INDICATION

NBAP

14. DCCH: RRC CONNECTION SETUP COMPLETE RRC


RRC

Huawei Confidential

Page 3

Mobile Terminated Call Setup Procedure (II) UE

Node B

RNC

MSC / VLR

15. DCCH: INITIAL DT [ PAGING RESPONSE ] RRC

MGW

16. SCCP CONNECTION RQ [ INITIAL UE MESSAGE [ PAGING RESPONSE ] ]

RRC SCCP

SCCP 17. SCCP CONNECTION CONFIRM

SCCP

SCCP

18. COMMON ID RANAP

RANAP

19. SECURITY MODE COMMAND RANAP

RRC RRC

20. SECURITY MODE COMMAND

21. SECURITY MODE COMPLETE

RANAP

RRC RRC

RANAP

22. SECURITY MODE COMPLETE

RANAP

23. DT [ SETUP ] RANAP

RRC

RRC

24. DCCH: DLDT [ SETUP ]

25. DCCH: ULDT [ CALL CONFIRMED ]

RANAP

RRC

RRC 26. DT [ CALL CONFIRMED ] RANAP

RANAP 27. BINDING ID, SPEECH CODE TYPE, B PARTY ROUTE

Here the RNC will perform a DRD decision and CAC decision for RAB


28. RAB ASSIGNMENT REQUEST RANAP

Huawei Confidential

RANAP

Page 4

Mobile Terminated Call Setup Procedure (III) UE

Node B

RNC

MSC / VLR

MGW

29. ESTABLISHMENT REQUEST ( AAL2 ) ALCAP

ALCAP 30. ESTABLISHMENT CONFIRM ( AAL2 )

ALCAP

NBAP

NBAP

ALCAP

31. RADIO LINK RECONFIG PREPARE

32. RADIO LINK RECONFIG READY

33. ESTABLISHMENT REQUEST (AAL2)

ALCAP

NBAP

NBAP

ALCAP

34. ESTABLISHMENT CONFIRM (AAL2) ALCAP

NBAP

ALCAP 35. RADIO LINK RECONFIG COMMIT

NBAP

36. DCCH: RADIO BEARER SETUP RRC

RRC 37. RAB ASSIGNMENT RESPONSE RANAP

RRC

38. DCCH: RADIO BEARER SETUP COMPLETE

RANAP

RRC

39. DCCH: ULDT [ ALERTING ] RRC

RRC 40. DT [ ALERTING ] RANAP

RANAP 41. ACM ISUP

42. DCCH: ULDT [ CONNECT ] RRC

RRC

Terminating UEs are considered to be in a call after CC Connect ACK

Originating UEs are considered to be in a call after CC Connect message

43. DT [ CONNECT ] RANAP

RANAP 44. OPEN CONNECTION 46. DT [ CONNECT ACK ]

RANAP

RANAP

RRC

ISUP

47. ANS (CONNECT)

46. DCCH: DLDT [ CONNECT ACK ] RRC


Huawei Confidential

Page 5

General Causes of failures (I)

• RF Reasons • Radio Parameter Problems • Miscellaneous causes


Huawei Confidential

Page 6

General Causes of failures - RF reasons (II)

 Poor DL coverage. The “fake coverage” phenomenon (the user sees the 3G icon on the screen in idle but cannot connect to any service). The cause could be overshooting cells but also excessive values of Qqualmin like -22 dB. Solution: Adjust the antenna azimuth and down tilt, add repeaters and RRUs, add micro cells. Any user should get a better signal than EcIo = -18 dB.

 Lack of Dominance (no clear Best server): Continuous change of best server leads to RRC failures and RAB failures.Solution: Establish a best server everywhere. Clear dominance.

 Poor UL coverage: The UE has not enough TX power to communicate with Node-B (even when there is low UL traffic on the cell). Solution: Adjust the antenna azimuth and down tilt, add repeaters, reduce CPICH power.

 Strong UL interference: Due to external interference or high UL traffic (the cell shrinking phenomenon). The UE will not be able to increase to more than 21 dBm for the preamble power and the RACH will fail - or synch will fail later. Solution: Up to operator‟s decision (implement more tilt ,CPICH power reduction, chase external source of interference or increase the number of Node-Bs to cope with traffic)

 Strong DL interference: Usually due to overshooting cell, external interference, high DL traffic on this cell and surrounding cells. The UE will miss the AI message for RACH and will fail to establish a call - or will fail to get synch in DL. Solution: Improve best server area (strong dominance)

 RF radiating system problems: 



Antenna‟s footprint not touching the ground properly: sites with over 120 m height and tilts around 3 degrees. More than 3/4 of the antenna pattern will not be touching the ground with a decent level of signal. Most calls are handled on side lobes. RF jumpers (feeding the antennas with RRU signal) are too long (should be no more than 3 meters, we‟ve seen cases in --- with 10 meters of ½” jumpers). This definitely leads to high noise factors and call setup failures. Also UL and DL coverage is very much limited.

 Missing neighbours: Leads to call setup failures due to poor signal. HUAWEI TECHNOLOGIES CO., LTD.

Huawei Confidential

Page 7



Poor DL coverage.

The “Fake coverage” phenomenon (user gets the 3G icon on his screen in idle but either cannot pass an RRC or a RAB). Cause is overshooting cells but also excessive values of Qqualmin like -22 dB. Solution: Adjust the antenna azimuth and down tilt, add repeaters and RRUs, add micro cells, improve best server, change Qqualmin. Any user should get a better signal than EcIo=-18 dB. If this level cannot be achieved it is better to display ” no service” on. user screen. User experience: 3G icon,3G signal User experience: 3G icon,3G signal

User experience: 3G icon, no 3G bar, no service accessibility. User „s perception: Very negative.

bar, good service accessibility. User „s perception: Positive.

bars, great service accessibility. User „s perception: Very Positive.

When -22>EcIo>-18 ; 80%>RRC_SR>20%

When -18>EcIo>-16 ; 95%>RRC_SR>80%

When -16>EcIo>-2 then RRC_SR>95%

Qqualmin

PRO

CONS

Comments

-22 dB

• User always see the 3G icon on his phone‟s screen (although it‟s a “fake” coverage the user does not always attempt to use the service) •Maximum traffic possible

• Bad customer experience but less NW signalling. •Not all call attempts are counted (not a clear perception of accessibility).

Will grab all extreme traffic leading quickly into DL Power congestion and accessibility issues.

-18 dB

• The user will not always have the 3G Icon on his phone‟s screen (but when icon is present service is 100% accessible) • Potential traffic decrease

• Great and real customer experience but increased signalling (coverage lost); • All “Call attempts” are counted (better performance perception of accessibility) due to this RRC_SR KPI may (or may not) be improved.

No more “fake coverage”. Decrease in DL Power Congestion.

-20 dB

Qqualmin=-20dB is suggested as a trade-off solution by Huawei.


Huawei Confidential

Page 8

General Causes of failures – Radio Parameter Problems(II) • Excessive values in object UCELLSELRESEL: Examples: Qqualmin<-18, IDLESINTRASEARCH ≠ 127, IDLEQHYST2S>1, IDLEQHYST1S>3.

• Improper settings of access parameters: No discrepancies found in UCELLACCESSRESTRICT • Inappropriate settings of preamble power ramp step and retransmission times: Current set of parameters is NOK (PREAMBLERETRANSMAX=20, CONSTANTVALUE-20, PowerRampStep=2, Mmax=8).

• Inappropriate setting of adjacent cells for UINTRAFREQNCELL: Qoffset1sn, Qoofset2sn out of the range (-4dB;+4dB). Wrong settings for Sintra (like 0 dB), Sinter( also like 0 dB).

• Inappropriate settings of synchronization parameters: Synch and Out-Of-Synch parameters for UL (N_INSYNC_IND=8, N_OUTSYNC_IND=8,and T_RLFAILURE=20), DL (T312=1, N312=1, N313=D20 ,T313=3 and N315=D20). Please remember that call re-establishment is activated for both UL and DL (great KPIs but acceptable user perception)

• Unsuitable power allocation rate for DL common channel:

No discrepancies found (PSCHPower, SSCHPower, BCHPower , MaxFachPower, PCHPower, AICHPowerOffset, PICHPowerOffset)

• Unsuitable initial power of uplink and downlink dedicated channel:

No discrepancies found for UL (DPCCH_Initial_Power = PCPICHPower - CPICH_RSCP + Uplink interference + DefaultConstantValue) and DL initial SIR target

• Unsuitable setting of uplink Initial SIR target value of dedicated channel: No discrepancies found for DL initial SIR Target

• Inappropriate setting of adjacent cells for UINTERFREQNCELL: • When 1900 and 850 MHz have significant azimuth difference why there is DRD just towards one 1900 cell and not for the other 1900 cell as well? • Why Qoffset1sn, Qoofset2sn are out of the range (-4;+4) on top of the IdleQhyst1s >2?


Huawei Confidential

Page 9

General Causes of failures –Miscellaneous causes(II)

•Transmission issues (fluctuating PATH, high BER, reduced capacity, routers down in the IP cloud). •Alarms on cells, on Node-Bs, on RNC, on transmission •Planning issues: traffic not properly shared between layers and NodeB, lack of a clear best server( no dominance), paging congestion due to LAC splitting issue. •Radio Congestion: • CE • DL Power • UL Power • R99 Codes • Iub bandwidth • SPU bottleneck • WMPT board bottleneck


Huawei Confidential

Page 10

How to identify and solve different issues?


Huawei Confidential

Page 11

RRC Access Failure Troubleshooting (I) Is cell/NodeB/RNC configuration the correct one?

YES

NO

Should be done daily (automatically and network wide) based on a defined template.

Where there any alarms on investigated cells (or any of it's neighbouring cells, intra or inter) ?

YES

NO

Every morning there should be an email with cells unavailable on previous day and duration of unavailability.

Is it a repetitive failure or a "one time" event?

YES

NO

If one time event, please wait one more day before to conclude. Could be a social event

If a repetitive failure (according to KPI values in the past) is it a slowly degradation (with traffic increase) or an event one (degraded seriously from a specific moment)?

YES

NO

If event one, go back to that day and see what was changed at that time and reconsider that change

Is the SHO factor less than 50%?

YES

NO

If not, please review its best server area, tilt, azimuth and CPICH power

Is this cell having full overlapping with other neighbouring cells? ( i.e. there's no direction user can move without having good coverage). Is any user, in any indoor environment within the footprint of this cell, able to get a decent RSCP and EcIo?

YES

NO

If no review your targeted coverage and accept current limitations and constraints due to location and/or number of Node-Bs.

Is the height of the antenna less than 100m?

YES

NO

If No, please do not expect a good RRC Success rate.

Is the total tilt of the cell more than 3 degree downtilt?

YES

NO

If no (and footprint is on a plain terrain) please take immediate actions to increase downtilt. Antenna RF pattern is hardly touching the ground, users are handled on side lobes. DL Power issues will occur.

is the cell Idle sintrasearch=127?

YES

NO

If no, please do not expect a good RRC Success rate.

is the cell idleQoffset1sn<4?

YES

NO

If no, please do not expect a good RRC Success rate with idleQoffset1sn>4dB

is the cell Idle idleQoffset2sn<2?

YES

NO

If no, please do not expect a good RRC Success rate with idleQoffset2sn>4dB

is the cell idleQhyst1<4?

YES

NO

If no, please do not expect a good RRC Success rate with idleQhyst1>4dB

is the cell idleQhyst2<2?

YES

NO

If no, please do not expect a good RRC Success rate with idleQhyst2>4dB


Huawei Confidential

Page 12

RRC Access Failure Troubleshooting (II) Date----

RNC1 RNC2 RNC3 RNC4

Sum of VS.RRC.AttConnEstab.Sum

6154003 7115377 5397822 1920647

Sum of Cell.RRC.Att.Fail

46702

73768

87275

12471

Sum of VS.RRC.Rej.Redir.Service

0

0

0

0

Sum of VS.RRC.Rej.ULIUBBand.Cong

0

0

0

0

Sum of VS.RRC.Rej.ULPower.Cong

14

10

0

0

Sum of VS.RRC.Rej.DLPower.Cong

135

118

1965

290

Sum of VS.RRC.Rej.DLIUBBand.Cong

0

0

0

0

Sum of VS.RRC.Rej.ULCE.Cong

1144

1352

721

507

Sum of VS.RRC.Rej.DLCE.Cong

822

11

0

0

Sum of VS.RRC.Rej.Code.Cong

12

41

372

0

Sum of VS.RRC.Rej.RL.Fail

30

50

419

0

Sum of VS.RRC.Rej.TNL.Fail

0

0

0

0

Sum of VS.RRC.FailConnEstab.Cong

2343

1552

3070

802

Sum of VS.RRC.Rej.Sum

2373

1602

3489

802

Sum of VS.RRC.SetupConnEstab Sum of VS.RRC.FailConnEstab.NoReply Sum of RRC.SuccConnEstab.sum

6151630 7113775 5394333 1919845 44006

71818

Conclusion: Most RRC failures (over 90%) are due to RRC no reply. For RRC issues, focus on overshooting cells first (to solve No reply), second on congested cells.

Huawei Confidential

11667

6107301 7041609 5310547 1908176

Here are most of the RRC failures occurring indicating poor UL coverage (overshooting)


83566

Page 13

RRC Access Failure Troubleshooting (III) Identify top N cells (more than 2000 RRC attempts per day and success rate is less than 98%)

Identify if RRC failures for a cell are due to SPU : (check ADD NODEB command to find the SPU for a cell/Node-B) .

YES

NO

SPU board is the issue when (VS.RRC.SuccConnEstabCPU / VS.RRC.AttConnEstabCPU) <97% (daily stats)

Solution: Open a ticket to TAC (should not happen after SPH226)

YES VS.RRC.Rej.ULIUBBand.Cong VS.RRC.Rej.DLIUBBand.Cong

Solution: On the RNC LMT, run the PING IP command to the IP of the NodeB during failing hours. If packet loss rate is greater than 0.1% please contact transmission engineers (to troubleshoot or upgrade)

VS.RRC.Rej.ULPower.Cong

Check configuration(ULTOTALEQUSERNUM=160; NBMULCACALGOSELSWITCH=ALGORITHM_SECOND). Solution: Increase ULTOTALEQUSERNUM to 180. If still failing: add carrier or reduce/balance traffic to other layers/cells.

VS.RRC.Rej.DLPower.Cong

Check configuration :UPCPICH and MAXTXPOWER should be at least 10 dB between the 2 values. DLOLCTRIGTHD=95.DL_UU_OLC=0. Solution1:add carrier or reduce/balance traffic to other layers/cells Solution2: change NBMDLCACALGOSELSWITCH=ALGORITHM_THIRD

Identify if more than 10% of the failures for a cell are due to congestion :

VS.RRC.Rej.ULCE.Cong

Solution per cell: change ULTTICREDITSFRESTHD from SF8 to 4SF4. CE resources for admission is in congestion status the new admitted 2ms HSUPA terminals will be mapped onto the 10ms TTI radio bearer. 2ms TTI HSUPA users whose bit rate is below the threshold of TTI reconfiguration(800Kbps) will be reconfigured to 10ms TTI Solution per RNC: decrease ULGBR from 32 to 16 for specific services

VS.RRC.Rej.DLCE.Cong

Solution per cell: ??? Solution per RNC: decrease DLGBR from 64 to 32 for specific services

VS.RRC.Rej.Code.Cong

Solution per cell is to activate Code reshuffling algo: in object UCELLALGOSWITCH change CELL_CODE_LDR to 1

NO

Identify if more than 10% of failures for a cell are due to RL.Failure :

VS.RRC.Rej.Rl.Fail

GO TO NEXT SLIDE


Huawei Confidential

Solution 1: If failures correlated with VS.IUB.FailRLSetup.NoReply then it is WMPT board congestion( add UTRP Board) Solution 2: run the PING IP command on the IP of the NodeB to detect congestion on the IuB. Solution 3: If failures correlated with RLM.FailRLSetupIub.Cong then it is WBBP board congestion

Page 14

RRC Access Failure Troubleshooting (IV) FROM PREVIOUS SLIDE

Identify if more than 10% of failures for a cell are due to TNL (Transport Network Layer) :

YES

VS.RRC.Rej.TNL.Cong

NO Identify if more than 10% of failures for a cell are due to FACH congestion :

YES

VS.RRC.AttConnEstab.Msg >> VS.RRC.AttConnEstab.Sum

Solution 1: Recheck configuration( IPPATHs of Nodeb has same capacity of transmission one;same for pair one) Solution 2: Run the PING IP command on the IP of the NodeB to detect congestion on the IuB.

Solution 1: (After SPH226) MOD UCELLALGOSWITCH: CellId=xxxxx, RsvdPara1=RSVDBIT5-1; (will improve CS success rate, will degrade PS success rate) Solution 2: Offload traffic

VS.CellFACHUEs>25

Solution: Offload traffic

VS.CRNCIubBytesFACH.Tx or VS.CRNCIubBytesPSR99.CCH.Tx are flat in time( limited)

Solution1: Offload traffic

FACH Channel utilisation>80%

Solution1: Offload traffic

NO

Identify if more than 10% of failures for a cell are due to VS.RRC.FailConnEstab.NoReply :

YES VS.MaxRTWP - VS.MeanRTWP > 10 dB

Check missing neighbours

NO If RRC Estab SR for whole RNC<99% and UU no reply is major cause

Solution 1: reduce HSUPA traffic Solution 2: Offload traffic Solution 3: Check external interference

Solution 1: Add important Neighbours Solution 2: Increase tilt to avoid risky overlaping footprints Solution 1 per RNC : modify RRC estab type to be on DCH SET URRCESTCAUSE:RRCCAUSE=TERMCAUSEUNKNOWN, SIGCHTYPE=DCH_3.4K_SIGNALLING, EFACHSWITCH=OFF; Take care as this will improve the RRC SR but will increase DL Power Solution 2 per RNC : stop SRB over HSDPA when DRD could be present SET UFRCCHLTYPEPARA, SRBCHLTYPE=HSUPA, SRBCHLTYPERRCEFFECTFLAG=TRUE

Please ask for HUAWEI support if after all those steps a cause for poor RRC failure could not be found


Huawei Confidential

Page 15

Top offending cell in xxx area with more than 10.000 RRC attempts per day: cell yyyyy •.

Data Sum of VS.RRC.AttConnEstab.Sum Sum of Cell.RRC.Att.Fail Sum of VS.RRC.Rej.Redir.Service Sum of VS.RRC.Rej.ULIUBBand.Cong Sum of VS.RRC.Rej.ULPower.Cong Sum of VS.RRC.Rej.DLPower.Cong Sum of VS.RRC.Rej.DLIUBBand.Cong Sum of VS.RRC.Rej.ULCE.Cong Sum of VS.RRC.Rej.DLCE.Cong Sum of VS.RRC.Rej.Code.Cong Sum of VS.RRC.Rej.RL.Fail Sum of VS.RRC.Rej.TNL.Fail Sum of VS.RRC.FailConnEstab.Cong Sum of VS.RRC.Rej.Sum Sum of VS.RRC.FailConnEstab.NoReply Sum of VS.RRC.SetupConnEstab Sum of RRC.SuccConnEstab.sum RRC_SR



Huawei Confidential

Page 16

cellID=yyyyy 73083 17110 0 0 0 163 0 0 0 0 0 0 163 163 16944 72920 55973 76.59%

2nd offending cell in xxx area with more than 10.000 RRC attempts per day: cell yyyyy Data cellID=yyyyy Sum of VS.RRC.AttConnEstab.Sum 75048 Sum of Cell.RRC.Att.Fail 5324 Sum of VS.RRC.Rej.Redir.Service 0 Sum of VS.RRC.Rej.ULIUBBand.Cong 0 Sum of VS.RRC.Rej.ULPower.Cong 0 Sum of VS.RRC.Rej.DLPower.Cong 115 Sum of VS.RRC.Rej.DLIUBBand.Cong 0 Sum of VS.RRC.Rej.ULCE.Cong 0 Sum of VS.RRC.Rej.DLCE.Cong 0 Sum of VS.RRC.Rej.Code.Cong 0 Sum of VS.RRC.Rej.RL.Fail 0 Sum of VS.RRC.Rej.TNL.Fail 0 Sum of VS.RRC.FailConnEstab.Cong 115 Sum of VS.RRC.Rej.Sum 115 Sum of VS.RRC.FailConnEstab.NoReply 5208 Sum of VS.RRC.SetupConnEstab 74933 Sum of RRC.SuccConnEstab.sum 69724 RRC_SR 92.91%

•.



Huawei Confidential

Page 17

3rd offending cell in xxx area with more than 10.000 RRC attempts per day: cell yyyyy •. Data Sum of VS.RRC.AttConnEstab.Sum Sum of Cell.RRC.Att.Fail Sum of VS.RRC.Rej.Redir.Service Sum of VS.RRC.Rej.ULIUBBand.Cong Sum of VS.RRC.Rej.ULPower.Cong Sum of VS.RRC.Rej.DLPower.Cong Sum of VS.RRC.Rej.DLIUBBand.Cong Sum of VS.RRC.Rej.ULCE.Cong Sum of VS.RRC.Rej.DLCE.Cong Sum of VS.RRC.Rej.Code.Cong Sum of VS.RRC.Rej.RL.Fail Sum of VS.RRC.Rej.TNL.Fail Sum of VS.RRC.FailConnEstab.Cong Sum of VS.RRC.Rej.Sum Sum of VS.RRC.FailConnEstab.NoReply Sum of VS.RRC.SetupConnEstab Sum of RRC.SuccConnEstab.sum RRC_SR Here are most of the RRC failures occurring indicating poor UL coverage (overshooting)


Huawei Confidential

Page 18

cellID=yyyyy 39512 2602 0 0 0 0 0 0 0 0 0 0 0 0 2602 39512 36910 93.41%

4th offending cell in xxxx area with more than 10.000 RRC attempts per day: cell yyyyy •. Data Sum of VS.RRC.AttConnEstab.Sum Sum of Cell.RRC.Att.Fail Sum of VS.RRC.Rej.Redir.Service



cellID=yyyyy 75162 3583 0

Sum of VS.RRC.Rej.ULIUBBand.Cong Sum of VS.RRC.Rej.ULPower.Cong Sum of VS.RRC.Rej.DLPower.Cong

0 0 439

Sum of VS.RRC.Rej.DLIUBBand.Cong Sum of VS.RRC.Rej.ULCE.Cong Sum of VS.RRC.Rej.DLCE.Cong Sum of VS.RRC.Rej.Code.Cong Sum of VS.RRC.Rej.RL.Fail Sum of VS.RRC.Rej.TNL.Fail Sum of VS.RRC.FailConnEstab.Cong Sum of VS.RRC.Rej.Sum

0 0 0 4 0 0 443 443

Sum of VS.RRC.FailConnEstab.NoReply Sum of VS.RRC.SetupConnEstab Sum of RRC.SuccConnEstab.sum RRC_SR

Huawei Confidential

Page 19

3137 74719 71579 95.23%

Paging Access Failure Troubleshooting-(I)case of one SCCPCH  2 Physical channels: – PICH (Paging Indicating Channel): This is just to inform the UE that it needs to initiate an RRC Connection request. Those are the details for this channel. – SCCPCH (Secondary Common Control Physical Channel). It carries paging messages themselves as well as packet messages for mobiles in cell FACH.


Huawei Confidential

Page 20

Paging Access Failure Troubleshooting-(II) -case of two SCCPCH  2 Physical channels: – PICH (Paging Indicating Channel): This is just to inform the UE that it needs to initiate an RRC Connection request. Those are the details for this channel.

2nd SCCPCH

– 2nd SCCPCH (Second Secondary Common Control Physical Channel). It carries only paging messages themselves.


Huawei Confidential

Page 21

Paging Access Failure Troubleshooting-(III) PICH timing in relation to P-CCPCH and S-CCPCH (extras from 3GPP 25.211-700) :

Paging ocassion

Paging message (3 IMSI or 5 TMSI)

Paging indicator


Huawei Confidential

Paging Access Failure Troubleshooting-(IV) PICH channel parameters: SF 256 used all the time.Each UE looks for a particular PICH timeslot according to several parameters broadcasted on SIBs : • PI number of paging indicators per radio frame. 3GPP allows values 18,38,72,144. It is broadcasted in Sysinfo5: PI-countperframe • SFN of the P-CCPCH where the PICH frame started. The SFN is known by UE immediately after synchronization with P-CCPCH. SFN range is from 0 to 4096. • K number of S-CCPCH and can be found in Sysinfo5 . Usually 1 or 2 ( same like in GSM combined or non-combined BCCH). • DRX cycle. UE will use the DRX=min (DRXPS,DRXCS). DRX cycle is broadcasted in Sysinfo1: cn-DRX-CycleLengthCoefficient (2 values broadcasted, one for each CN domain) • IMSI known from U-SIM. • Frame offset =Ts-ccpch,k –Tpich (see previous slide). Ts-ccpch,k = Tk  256 chip, Tk  {0, 1, …, 149} and can be found in Sysinfo5: and it is called timming offset. For particular UTRAN timming offset=0(S-CCPCH and P-CCPCH are time aligned). Tpich = 7680 chips as a fix value forced by 3GPP. • A paging indicator set to “1” indicates that the UE should read the S-CCPCH of the corresponding frame. • Total number of chips in one 10msec radio frame is 38400. PICH channel can transmit (38400/256) 150 indicator modulation symbols or (150X2) 300 bits. Only the first 288 of these are used, leaving the last 12 bits undefined • More details in 3GPP specs: 25.211-700 and in 25.331-710 RRC protocol specification

•

PO= {(IMSI div K) mod (DRX cycle length div PBP)} * PBP + n * DRX cycle length + Frame Offset Where n = 0,1,2… as long as SFN is below its maximum value ,for FDD PBP=1

PI=(IMSI/8192) mod36


Huawei Confidential

•

PI = DRX Index mod Np Where DRX Index = IMSI div 8192 If we consider particular settings: DRXcycle=7=>128 frames Frame offset =-7860 chips PBP=1 K=1 (there„s only one S-CCPCH that carries PCH) PI=Np=36 PO= (IMSI)mod128+ n* 128 -7860 chips

Paging Access Failure Troubleshooting-(V) PICH frame structure : • A group of bi=1 means there‟s a paging and UE should read it‟s very first paging occasion. • A group of bi=0 means there‟s no paging and UE could go back to idle till next paging indicators.

•

• •

More bits inside a PI means a greater probability to decode the paging indicator but less capacity of the paging channel and power consumption for UE. Less bits means a lower probability for the UE to decode the paging indicator but longer battery life of the UE. Best solution is a mid-way one: PI=36.


Huawei Confidential

Paging Access Failure Troubleshooting-(VII)

•

From all this information what do you need to know?: • there can be several places where paging could get congested: Iu interface, IuB interface, RNC boards, or PCH interface . PICH channel is the only channel that is never congested! • Check with CN how many paging repetitions have, how do they page: by IMSI or by TMSI. If first paging fails how many repetitions? Last paging is network wide or LAC wide only? • --- is currently facing PCH channel load: all smart phones are in cell PCH state. In this state can only receive paging but can not transmit any data. Any paging for a UE it is sent specifically to that cell. How RNC knows where is such an UE? By cell update!. Every time UE changes the cell in cell PCH there is a cell update+cellupdate confirm, utran mobility information confirm. That means that the RNC is aware about new location of the UE. • How much is the paging success now in --- network? • What solutions we have to offload the PCH channel?: • LAC split. • Page by TMSI • Reduce ping-pongs (and reselections) • Improve best server area and reduce overlapping


Huawei Confidential

RACH Access Failure Troubleshooting (I) •Why are RACH parameter VERY important? Because it impacts strongly user experience (also called E2E=end-to-end user experience) No performance indicators. Only estimation by RTWP, load of the RACH channel etc..

Enough performance counters


Huawei Confidential

Page 26

RACH Access Failure Troubleshooting (II) The answer on AICH must be a specific positive response for the specific RACH sent Max_TX_power_on_PRACH NB01

Power_Ramp_Step :

….

Preamble_Initial_Power :

Uplink/UE/PRACH

Preamble 1

Pp-m :

….

Message part

Preamble n AICH_Transmission_Timing

Preamble_Retrans_Max :

MMax Parameters for RACH/PRACH: •NBO1( 0 NBO1min  NBO1  NBO1max ) is the time between 2 ramping power of the preamble within the same preamble cycle. •Preamble_Retrans_Max is the maximum number of preamble that can be sent in a cycle. •Mmax is the maximum number of preamble cycles. •Preamble_Initial_Power = Primary CPICH TX power – CPICH_RSCP + UL interference + Constant Value •Constant value is an initial value to start the first preamble power usually is -24. •UL interference is the latest value broadcasted by the NodeB in SIB7. Ue needs to decode this value before being able to transmit RACH. •Power_Ramp_Step is the how much the preamble power should be increased after each No ack received on AICH. •Power offset P p-m = Pmessage-control – Ppreamble, measured in dB, between the power of the last transmitted preamble and the control part of the random-access message. •AICH_Transmission_Timing is the time when the RACH message must be transmitted after positive AICH was received( there are other parameters too) RACH is a common type transport channel in the uplink. RACHs are always mapped one-to-one onto physical channels (PRACHs), i.e. there is no physical layer multiplexing of RACHs, and there can only be one RACH TrCH and no other TrCH in a RACH CCTrCH. Service multiplexing is handled by the MAC layer. In one cell several RACHs/PRACHs may be configured. If more than one PRACH is configured in a cell, the UE performs PRACH selection RACH message mandatory parameters: -UE identity( IMSI,IMSI+LAI, TMSI, IMEI-when no USIM is inserted) -RRC establishment cause (31 causes) -radio bearer ID( AS or NAS, UM or TM or AM) -release5 indicator -measurements results on RACH(like EcNo of the serving cell).


Huawei Confidential

Page 27

RACH Access Failure Troubleshooting-(III) From all this information what do you need to know?: • Current RACH parameters are not optimal: allows the UE to increase the power 20 dBm more than the RTWP(CONSTANTVALUE=-20, PREAMBLERETRANSMAX=20, POWERRAMPSTEP=2). Due to this RTWP increase, due to this RACH increases and so on(it creates an avalanche effect). Better have longer call setup time for one UE (RACH failures due to missing NB relations of overshooting cells) instead of having entire cell shrinked due to one UE not being able to transmit RACH message.

•

•

Missing neighbours, lack of best server area and poor UL coverage influence a lot the RACH success rate.

•

Cell radius is now at 29.000 km. Make sure there are no UE from a larger distance(path distance) else will fail on RACH.

•

Spreaders inside the Node-B are limited. Multipath ( long distance) is not good for resource consumptions and so RACH messages might be missed.


Huawei Confidential

Preliminary conclusions •

Most attempt failures are related to planning

•

Plenty of attempts failures not recorded within the performance file (When EcIo is worse than -18 very few RACH “reach” the Node-Bs)


Huawei Confidential

Page 29

Thank you www.huawei.com

Access Failures Troubleshooting Workshop

Recommend Documents