LOW HSDPA Throughput_Huawei NodeB

[miles12]

Dear Expert

We are facing low throughput for below nodeB. We have BW choking as well. But transmission guys said that there is no Tx issue from there end and they set in 30Mbps. where as its choking in near 6mbps.

They said that, nodeB might have Radio configuration issue. Please help me to fix the problem.




BR and TIA
Miles

[electron]

Dear Expert

We are facing low throughput for below nodeB. We have BW choking as well. But transmission guys said that there is no Tx issue from there end and they set in 30Mbps. where as its choking in near 6mbps.

They said that, nodeB might have Radio configuration issue. Please help me to fix the problem.




BR and TIA
Miles

Hi,

As a first step lock all HSUPA channels on the site and make HSDPA test and share the result.If HS Throughput improve for sure there is transport issue and the cause is HSDPA flow control causing the issue . After demonstration of mentioned activity will explain further where to find the root cause


B
R

[babak1349]

Dear Expert

We are facing low throughput for below nodeB. We have BW choking as well. But transmission guys said that there is no Tx issue from there end and they set in 30Mbps. where as its choking in near 6mbps.

They said that, nodeB might have Radio configuration issue. Please help me to fix the problem.




BR and TIA
Miles

Dear Miles,
I think the first thing is to change your mindset about trouble shooting....just sending a graph will not give enough information to help you out..
To trouble shoot :
1) understand the phenomena and validate it ( So do a DT during day and night) Doing DT at night help to evaluate the low traffic/load situation
2) if you have suspicious about TN, you can ask TN team to use a packet generator to evaluate the real available bandwidth ( it is physical test )
3) understanding what limits the throughput , will help to locate the issue for example how many HSDPA user you have ? Do you have limitation on user profile ? How much is code utilization ?How much is power utilization ? How about CQI, how it is ? do you have problem with HSDPA/HSUPA user or you have problem with HSDPA/DCH users or both ? and so on....
4) if you increase Static allocated SF16 for HSDPA , does the HSDPA user throughput will increase ?( please be careful it might cause code blocking for R99 so you need to do the test in low load situation)
5) do you have any parameter in nodeB which limits flow control of HSDPA and its throughput ?
6) how is TN setting and QOS implementation in nodeB and TN?
7) does the TN sharesd with other technologies ? is it in a TN chain ? if it is on the same TN chain with the other nodeBs, do they have similar problem ? if they do so most likely you have TN issue ?

So an other questions might be answered .. to find the root cause ?

Now what I understood from the chart.. I can not see the hours .. so in the early morning or in the low load situation the throughput increases ?
it is a KPI from TN or IP interface of nodeB ? how about the HSDPA throughput from the cells ? do you see chopped throughput too ? do you see different maximum throughput of cell HSDPA throughput ?

hope those comments help you

cheers

[samuraial]

Continuing the babak1349 comments that you should follow to the letter you can check also by doing a test on a number you already have in site or by just doing a test on a smartphone that is attached to the current cell you want to see the throughput.

It is called an autoping test where the user is being sent packet to simulate a full load.
Right click on your RNC in U2000 at topology and go at "maintenance client", go at Trace -> UE trace.
enter the IMSI of the UE you are tracing, click debug mode -> go at other tab and click auto-ping, enter the packet length 1428 or 1400.
start the trace.
Now go at "Monitor" tab and click connection performance monitoring. select DL throughput after entering the same IMSI and start it.
You can also check the EcNp and RSCP with a new connection performance monitoring so you can see if the user is in good radio condition.
This way if you do the test in good radio condition, single user in the cell you can see the graph and do the analyze.
If you do it on a cell where you are the only user and good radio conditions and your max throughput is limited you have problems in transmission, but also check your RTWP in the network as high RTWP can affect your throughput.
don't forget to put DPA codes min 1 and max 14 and automatic allocate them.


B/R

[babak1349]

Continuing the babak1349 comments that you should follow to the letter you can check also by doing a test on a number you already have in site or by just doing a test on a smartphone that is attached to the current cell you want to see the throughput.

It is called an autoping test where the user is being sent packet to simulate a full load.
Right click on your RNC in U2000 at topology and go at "maintenance client", go at Trace -> UE trace.
enter the IMSI of the UE you are tracing, click debug mode -> go at other tab and click auto-ping, enter the packet length 1428 or 1400.
start the trace.
Now go at "Monitor" tab and click connection performance monitoring. select DL throughput after entering the same IMSI and start it.
You can also check the EcNp and RSCP with a new connection performance monitoring so you can see if the user is in good radio condition.
This way if you do the test in good radio condition, single user in the cell you can see the graph and do the analyze.
If you do it on a cell where you are the only user and good radio conditions and your max throughput is limited you have problems in transmission, but also check your RTWP in the network as high RTWP can affect your throughput.
don't forget to put DPA codes min 1 and max 14 and automatic allocate them.


B/R

hello Mate
I heard about that cool feature in Huawei RNC ... does it use UDP or TCP mode ?

cheers
ا

[samuraial]

Hello M8,

If you are speaking about IUPS tests you can choose UPD or TCP in that case.
In this case it is the same UPD test to check the radio capacity.

B/R


hello Mate


I heard about that cool feature in Huawei RNC ... does it use UDP or TCP mode ?

cheers
ا

[miles12]

Dear Babak
TN team replied us , nodB has no BW problem. its set with 30Mbps. Any way I will check your point. Thanks a lot for your nice description.


BR/
Miles

[miles12]

Dear samuraial

Thanks for your great support. Unfortunately Maintenance client not working. I will share you my findings after problem fix.



Continuing the babak1349 comments that you should follow to the letter you can check also by doing a test on a number you already have in site or by just doing a test on a smartphone that is attached to the current cell you want to see the throughput.

It is called an autoping test where the user is being sent packet to simulate a full load.
Right click on your RNC in U2000 at topology and go at "maintenance client", go at Trace -> UE trace.
enter the IMSI of the UE you are tracing, click debug mode -> go at other tab and click auto-ping, enter the packet length 1428 or 1400.
start the trace.
Now go at "Monitor" tab and click connection performance monitoring. select DL throughput after entering the same IMSI and start it.
You can also check the EcNp and RSCP with a new connection performance monitoring so you can see if the user is in good radio condition.
This way if you do the test in good radio condition, single user in the cell you can see the graph and do the analyze.
If you do it on a cell where you are the only user and good radio conditions and your max throughput is limited you have problems in transmission, but also check your RTWP in the network as high RTWP can affect your throughput.
don't forget to put DPA codes min 1 and max 14 and automatic allocate them.


B/R

[paulus]

hi Miles,

Have u check IPPMDropMeans (Max and Average). if it >2%, or even >10%, that means transport problem happen. Congestion transport may occur.

Check also, is there active alarm of IPClock link failure, IPPMActivation failure?

thnks
Br//Paul

[tremalnike]

Hi,
you have to check also how many HS codes are configured and used.
May be issue is due to configuration problem.