3G Uplink Congestion

[RFSTUD]

Hi everyone,
I am a new staff faced with a major problem. The operator I am working for, is currently experiencing uu uplink congestion on the 3G network. Therefore, it is very hard for calls to get admission. I will be much grateful if you can point me to some efforts of reducing this congestion. We are currently using 128 CE for the cell and using Huawei eqeuipment. Thanks a lot for your help!

[jamal]

Dear Mate,

Are U already check UL Interference?
Hopefully helping U.

BR/
JM

[RFSTUD]

We already down tilted the antenna by 2 degrees to reduce the load; but the problem still persists. I dont really know how to check for uplink interference- do I need specialsed equipment like spectrum analyser or can I get it from the NodeB LMT? Any resources how to carry it out?

[jamal]

U can check uplink interference from RNC measurement.
Use moshell, input command "cedh" then look for the cell related at the ulInt column.

More simple use command "cedr -r <iublink name>.

CMIIW, for detail please look into AMOS or MOSHELL help page.

After, make sure there are an UL interference then U can search using spectrum analyzer. Usually, there is some pico repeater installed near the RBS.

Goog luck mate.

[RFSTUD]

I have searched cedh or cedr but could not find it on the Huawei BSC 6810. Appreciate your help. I'm still searching for solutions.

[prasath085]

just check the RTWP for adjacent cell and also check any 2g repeater is placed near the site

[jamal]

Im' sorry mate, I didn't realized that you are using Huawei RNC. I thought U are using e*******.

Sorry mate, that's all I know.

[agenov]

Hi everyone,
I am a new staff faced with a major problem. The operator I am working for, is currently experiencing uu uplink congestion on the 3G network. Therefore, it is very hard for calls to get admission. I will be much grateful if you can point me to some efforts of reducing this congestion. We are currently using 128 CE for the cell and using Huawei eqeuipment. Thanks a lot for your help!


First check which exactly resource can not pass admission. Could be NodeB channel elements or ENU in the UL and etc. there are statistics in M2000 for that.
Then you may try to increase problematic capacity or to decrease load by different algorithms for that just like LDR for example.... Let us know what your findings are in order to get further.

BR
Alex

[zhanglw268]

Hi everyone,
I am a new staff faced with a major problem. The operator I am working for, is currently experiencing uu uplink congestion on the 3G network. Therefore, it is very hard for calls to get admission. I will be much grateful if you can point me to some efforts of reducing this congestion. We are currently using 128 CE for the cell and using Huawei eqeuipment. Thanks a lot for your help!

I assume that you have 128 UL CEs.
1. check what's the congestion threshold setting, if it was set to low, you can increase the threshold value
2. If congestion are perceived at RRC level. Please consider to change the triggered threshold. It can be set on Call Admission Control algorithm (CAC). Using high thresholds allows efficient CE resources utilization.
3. HSUPA share with R99's CEs but HSDPA won't, check whether you have HSUPA feature enabled and used by your customer, you can disable this feature if your customer don't purchase it
4. check whether you have DCCC feature disabled, if it is, please enable it.

For example, Several parameters can be considered.
Decreasing UL InitalBitRate to be ex: D32, this would enhance general accessibility

Enable UL & DL CE LDR:codeadj, CSINTERRAT, BE Rate Reduction

changed the ulmidratethd from 128 to 64 and dcccratethd to from 64 to 32

Regards

Zhanglw268

[zee86]

buddy its looking like parameter setting problem
check when this problem start and try to find out what change had been done that time

[hari_2208]

i would like to add one more thing...

pls check the rach parameter are properly defined else MS will keep on trying & creating UL interference.

Also ensure your VSWR in case you are using feeder cables is good,as issue with the cables or in the uplink circuits internal to the bts will cause poor uplink ...leading to cascading effects.

also check if the issue is seen only on busy hour or almost entire time,this will also help you isolate the issue due to actual load or possible interference.

Its also adviced to check the iub interface health/util

[sonerss]

Hi jamal

Can you share us most using moshell command list??

Thanks..


U can check uplink interference from RNC measurement.
Use moshell, input command "cedh" then look for the cell related at the ulInt column.

More simple use command "cedr -r <iublink name>.

CMIIW, for detail please look into AMOS or MOSHELL help page.

After, make sure there are an UL interference then U can search using spectrum analyzer. Usually, there is some pico repeater installed near the RBS.

Goog luck mate.

[agenov]

i would like to add one more thing...

pls check the rach parameter are properly defined else MS will keep on trying & creating UL interference.

Also ensure your VSWR in case you are using feeder cables is good,as issue with the cables or in the uplink circuits internal to the bts will cause poor uplink ...leading to cascading effects.

also check if the issue is seen only on busy hour or almost entire time,this will also help you isolate the issue due to actual load or possible interference.

Its also adviced to check the iub interface health/util


I don't believe high RTWP can cause cell congestion in the UL and CAC rejections. By default Huawei use algorithm 2 for CAC and is not recommendable to change it, at least till RAN12 as far as i know. CAC by Iub and Codes are mandatory, and optionally are UL/DL resources so he has to check them at firstly, depending on their settings of course.

[RFSTUD]

Thanks Prasath085. Your comments are really appreciated.The latest readings for the RTWP for the adjacent cell is -102.7dBm. I suppose that is not too bad. There is no 2G repeater near the site. We are planning to do a DT to study the radio propagation characteristics but I doubt its an interference issue. My assumption is that its a radio congestion issue.

[RFSTUD]

Im' sorry mate, I didn't realized that you are using Huawei RNC. I thought U are using e*******.

Sorry mate, that's all I know.

No worries bro, Thanks for your help!

[RFSTUD]

First check which exactly resource can not pass admission. Could be NodeB channel elements or ENU in the UL and etc. there are statistics in M2000 for that.
Then you may try to increase problematic capacity or to decrease load by different algorithms for that just like LDR for example.... Let us know what your findings are in order to get further.

BR
Alex

Thanks Alex. I'll check the stats tomorrow and will get back to you. You think its CE resource limitation? Load reshuffling might be a good idea if it is capacity issue on a single cell but I'm not sure how to implement it.

[RFSTUD]

I assume that you have 128 UL CEs.
1. check what's the congestion threshold setting, if it was set to low, you can increase the threshold value
2. If congestion are perceived at RRC level. Please consider to change the triggered threshold. It can be set on Call Admission Control algorithm (CAC). Using high thresholds allows efficient CE resources utilization.
3. HSUPA share with R99's CEs but HSDPA won't, check whether you have HSUPA feature enabled and used by your customer, you can disable this feature if your customer don't purchase it
4. check whether you have DCCC feature disabled, if it is, please enable it.

For example, Several parameters can be considered.
Decreasing UL InitalBitRate to be ex: D32, this would enhance general accessibility

Enable UL & DL CE LDR:codeadj, CSINTERRAT, BE Rate Reduction

changed the ulmidratethd from 128 to 64 and dcccratethd to from 64 to 32

Regards

Zhanglw268

Mighty thanks Zhanglw268 for your comprehensive input.
1. UL congestion threshold is set at 80 for AMR services. which is my highest priority. The threshold for handover is set at 85.
2.Tried configuring it through CAC but needed to increase the UL handover thd for AMR in order to increase the congestion threshold. I do not know how to do that. I'm a little bit hesitant touching the handover threshold too as it might be tied to other parameters.
3. HSUPA is disabled. Only HSDPA is enabled.
4. DCCC is enabled. UL InitalBitRate is D32. ulmidratethd is 128 but dcccratethd is 32.

Lastly, what effect will enabling "UL & DL CE LDR:codeadj, CSINTERRAT, BE Rate Reduction" have. Is it a load reshuffling parameter?

[RFSTUD]

buddy its looking like parameter setting problem
check when this problem start and try to find out what change had been done that time

Thanks Zee86, apparently, the problem has been around for sometime. Before I even started working for the operator. However, the problem is now very critical because its very difficult to initiate 3G calls.

[RFSTUD]

i would like to add one more thing...

pls check the rach parameter are properly defined else MS will keep on trying & creating UL interference.

Also ensure your VSWR in case you are using feeder cables is good,as issue with the cables or in the uplink circuits internal to the bts will cause poor uplink ...leading to cascading effects.

also check if the issue is seen only on busy hour or almost entire time,this will also help you isolate the issue due to actual load or possible interference.

Its also adviced to check the iub interface health/util


Good advice bro. I will check the VSWR tomorrow and also the RACH parameter. The problem is worst during busy hour.
In addition, I will also check the iub interface and get back to you.

Thanks a lot guys for your invaluable help. Its truly truly appreciated.

[adedia]

128 UL CE for Huawei are just too small. If the value are correct. You must upgrade.

[agenov]

Thanks Alex. I'll check the stats tomorrow and will get back to you. You think its CE resource limitation? Load reshuffling might be a good idea if it is capacity issue on a single cell but I'm not sure how to implement it.


If you can not make a call almost at 100% this is CE congestion. What you can do initially is to enabled HSUPA, activate DCCC for HSUPA as it was proposed already (set different steps appropriate with your strategy) as well as introduction of "UE state transition" (Cell_Fach-PCH and etc.) functionality that will give immediate effect in regards to the all dedicated resources/licenses. What you can do by LDR is "BE rate reduction" that will allow you to reduce throughput at the price of CEs.Evaluate that NodeB for second base band unit expansion followed by CEs increasement after all above.

Cheers,
Alex

[sala70vic]

Mighty thanks Zhanglw268 for your comprehensive input.
1. UL congestion threshold is set at 80 for AMR services. which is my highest priority. The threshold for handover is set at 85.
2.Tried configuring it through CAC but needed to increase the UL handover thd for AMR in order to increase the congestion threshold. I do not know how to do that. I'm a little bit hesitant touching the handover threshold too as it might be tied to other parameters.
3. HSUPA is disabled. Only HSDPA is enabled.
4. DCCC is enabled. UL InitalBitRate is D32. ulmidratethd is 128 but dcccratethd is 32.

Lastly, what effect will enabling "UL & DL CE LDR:codeadj, CSINTERRAT, BE Rate Reduction" have. Is it a load reshuffling parameter?



Hi,

Yes this is the load reshuffling algorithm, In addition to the above advices:
1- Try to reduce the HO area, check your SHO overheads to minimize using more CE during SHO
2- you mentioned that 128 CE for cell, CEs are shared for all NodeBs' cells, do you have only on cell in this NodeB?
3- Analyse the traffic in busy hour and try to analyse the traffic during this hour in order to be know which service is consuming your CEs?

Hope this helps, regards.

[patek1968]

Here in our country,we are using 384CE per Cell in UL but still congested..

[zee86]

Here in our country,we are using 384CE per Cell in UL but still congested..

bro check at which time congestion come
and request for one more TXB card expension
in the mean time sf4ul and sf8ul 0
this will reduse your congetion :)

[RFSTUD]

Thanks for your help Zee; we did CE board expansion and the problem is resolved. cheers

[RFSTUD]

128 UL CE for Huawei are just too small. If the value are correct. You must upgrade.
you are absolutely right Adedia, 128 UL CE PER NODEB is too small....we have now upgraded to 256 ul+dl CE and there is no more congestion.

[RFSTUD]

128 UL CE for Huawei are just too small. If the value are correct. You must upgrade.


If you can not make a call almost at 100% this is CE congestion. What you can do initially is to enabled HSUPA, activate DCCC for HSUPA as it was proposed already (set different steps appropriate with your strategy) as well as introduction of "UE state transition" (Cell_Fach-PCH and etc.) functionality that will give immediate effect in regards to the all dedicated resources/licenses. What you can do by LDR is "BE rate reduction" that will allow you to reduce throughput at the price of CEs.Evaluate that NodeB for second base band unit expansion followed by CEs increasement after all above.

Cheers,
Alex

Thanks Alex for your invaluable help! The problem was definitely the physical CE or its lack thereof.
Initially we were using on average 96 ul+dl CE per nodeB. We then increased all NodeBs to 128 ul+dl CE. After this, the KPI: VS.RRC.Rej.UL.CE.Cong(times) REDUCED from the thousands (for the busiest sites) to practically 0. However, the KPI: VS.RRC.Rej.RL.Fail(times) was still in the thousands for the busiest sites. We have just upgraded the CE for these sites to 256. Now, we are not experiencing any congestion for both voice and data.

I will now optimize and implement the state transitions and set the BE rate reduction .

Thanks again!

[theumar@hotmail.com]

Thanks for info sharing, Can you please help with following

1. how to allocate CE? is it common resource pool or it is user based assignment dedicated for different services?
2. maximum how many CEs at one NodeB can be assigned? and what is the usual values of CEs
3. What is the meaning of congestion perceived at RNC level? and by changing it to Call Admission Control Algorithm how it will helpful in reducing congestion?

[dNoB]

1) Number of CEs allocated according to the service being requested.
2) you can have as many CEs at one NodeB as long as you have the licenses and your hardware supports the expansion. Different vendors have NodeBs with different capacities. Hope the experts can advise re this, hardware model and capacities.
3) RNC level congestions may include link congestions; need to set levels of congestions to trigger network warnings/alerts, in time for prompt resolutions.

[Bronco30]

Chek how many ENU (CELLCAC) is configured in the cell, and increase them.
BR

[Redrodri]

According to the version you have, (ULtotalequivalentusernumber) are configured by default to 80 or 95, changes to higher values​​, the maximum is 200.