PDA

View Full Version : SDCCH Fail (abrupt discontinuous LU ATTEMPT Increase)



asifhaider919
2011-11-12, 09:21 AM
I am working NSN S14, with below problem.

Sudden increase of SDCCH REQ, the increased cause is SDCCH_LU_ATTEMPT, SDCCH resources are assigned and then there is no established indication. T3101 expires so I have High SDCCH ABIS FAIL

The problem is random and starts mostly after midnight and automatically resolves after an average of 6-8 hours.

The sites are not on LA border, are present in same BSC and Same MSC.

There is no specific pattern of occurence of this behaviour, it random will appear on any site. I have dual BCCH network the colocated cells are impacted usually together but not always. But geographically the problem is not centered to some particular area, its spread.

There is no problem of C0-BCCH/BSIC

This cannot be a GHOST access as the casue is LU attempt and, GHOST cannot always be an LU but a random distribution over all causes, counter GHOST CCCH is nearly equal to zero.

Also one thing noticed is that although the problem resolves automatically after few hours, if the sector is reset it resolves at the same time.

Nothing could be observed during the drive test, there were no failures in the drive test.


I just surfed the web and saw some people of other vendors talking about the same problem but with no conclusion.


There are no RAJ Rejections, no AGCH overload, no channel activation failures, no Busy Attempt

The difference between SDCCH_LU_ATTEMPT - SDCCH_LOC_UPD = nearly equal to T3101 Timer Expiry

SDCCH_LU_ATTEMPT = no of SD request with casue location update
SDCCH_LOC_UPD = no of successful LOC update

an example case:

SDCCH_REQ = 1000
SDCCH_SEIZURE_ATTEMPTS =1000

SDCCH_LU_ATTEMPT = 900
SDCCH_LOC_UPD = 50
SDCCH_LU_ATTEMPT_FAIL=0

SDCCH ABIS FAIL CALL = 850
T3101 Expiry = 850


Expecting some good replys :)

plannerguy
2011-11-19, 08:45 PM
which vendor are u using....

also is the sectors facing any railway line or busy NH corridors..


regards
Plannerguy

asifhaider919
2011-11-19, 09:43 PM
We have identified the problem, it was due to faulty MS, probably chinese versions randomly sending repetitive LU update attempts on consecutive timeslots. confirmed from Abis traces.

plannerguy
2011-11-19, 09:46 PM
Hi

So finally i guess your problem is resolved:D

Regards
Plannerguy

louvre
2011-11-20, 03:39 AM
We have identified the problem, it was due to faulty MS, probably chinese versions randomly sending repetitive LU update attempts on consecutive timeslots. confirmed from Abis traces.

Hi,

Could you provide me the make of the mobile i would like to do some test on a network, where i have seen some weird stuff happening!!.

asifhaider919
2011-11-20, 05:20 AM
Hi,

Could you provide me the make of the mobile i would like to do some test on a network, where i have seen some weird stuff happening!!.

The make of the mobile couldnot be identified as during the RACH access no info of mobile is transferred. We are going to have a patch soon to solve this fake problem using SDCCH resource allocation.

blackhole
2011-11-23, 06:52 PM
The make of the mobile couldnot be identified as during the RACH access no info of mobile is transferred. We are going to have a patch soon to solve this fake problem using SDCCH resource allocation.

Hi asifhaider919 (http://www.finetopix.com/member.php?15252-asifhaider919),
Is your vendor Ecrisson? I hear that Ecrisson have a patch to prevent SDDCH congestion from crazy mobiles like this but still confuse how to prevent it. In fact, the BTS cannot know where the RACH request come from? How to prevent duplicate requests? Can you share the idea?

Blackhole
http://www.finetopix.com/images/statusicon/user-offline.png

blackhole
2011-11-23, 06:55 PM
Sorry, I don't read this thread from the begining. So you work for NSN :D Anyway, can you share the idea?

zhanglw268
2011-11-24, 05:26 AM
change T3101 Expiry to 5 seconds will make your problem almost disappeared most of time.

m121
2011-11-25, 08:51 PM
I dont't think that Setting T3101 to 5s will improve droprate it probably will only worsen SDCCH congestion because resources will be allocated for longer period of time.
Sudden spikes in SDCCH assignments nowadays are expierenced by almoust all operators and all vendors and I haven't heard that there is any solution to this problem.
To improve droprate and ease congestion I think that only way to help is to give to NOC a tool for online statistics monitoring so they can reset cell if high SDCCH assignment failure rate is in any cell.

melkaiy
2011-11-25, 09:42 PM
your problem is in the paging capacity in the bss side. Good location area update is once in every 60 minutes that is composed of 5 sdcch with 2% blocking probability.

if this one helps please add reps if you like it. thnx