NSN problem of 0 kbps on HSDPA

[josef]

Dear borthers,

In our network, we have many complain about throughput on HSDPA. The problem is that throughput went many times from good value to 0 kbps and it stay on this value. After restarting the session, the same problem appear.

What can be the reasons of this problem? If the problem is related to radio, Is there any radio parameters related to this issue that we can test?

Thanks for your help :)

Josef.

[elkador]

Vendor, release, technology deployed?
Any specific handset impacted, or all of them?

[josef]

Vendor, release, technology deployed?
Any specific handset impacted, or all of them?

Vendor: NSN
Release: RU10
Technology: HSDPA/ HSUPA U900 & U2100

all the handsets are impacted.

Thanks bro

Josef

[redrooster]

Need more information :)

Is this on just one cell or more? Is it constant? In addition, what is the voice traffic load at the time? Obviously if all the "codes" are being used for voice none are available for HSDPA (8 voice calls equates to one hsdpa code).

Cheers

Redrooster

[EJAY]

Hi Josef,
THere are several reasons for that problem.
- number of users
- capacity- ce iub, etc
- sgsn/ core problem
check first all ur parameter. WBTS parameter and RNC parameter.

Regards,
EJAY

[ash_or]

Hi josef..
have you try to check on core/sgsn/ggsn
maybe there some limitation

[s52d]

Hi Josef,
THere are several reasons for that problem.
- number of users
- capacity- ce iub, etc
- sgsn/ core problem
check first all ur parameter. WBTS parameter and RNC parameter.

Regards,
EJAY

How about IuB? If signalling is OK, but no user plane, you get happy 0 kbit/sec users.
And terminals do not want to pick another cell: signalling is OK!

This is common with dual stack IuB: signalling via ATM (normally works),
and user plane via IP, where some IT people never learned how NOT to mess with
routing etc....

BR
s52d

[syeddawer]

1. First of all, since ur network is deployed and u do get down/upload of data, i dont guess there is a problem with the user plane.

2. Radio: You need to check at what instances this data outage occurs. Which cells are impacted. what is the traffic trend and network utilization shares of these cells. Also check AC statistics.

3. Packet Core: You need to check the utilization of Gi/Gn Interfaces (Gi specially). I strongly suspect somewhere on ur Packet Core, a network element is dropping packets, being heavily oveutilized. It can be the firewall even. Your routers/switches can be hitting 100% utilization, hence the packet drop.

Hope this helps.

Br..
D

[josef]

Need more information :)

Is this on just one cell or more? Is it constant? In addition, what is the voice traffic load at the time? Obviously if all the "codes" are being used for voice none are available for HSDPA (8 voice calls equates to one hsdpa code).

Cheers

Redrooster

This problem appear in many cells. but You can not produce it in some nodeBs. for voice traffic, we don't have so much in 3G. mainly 2G voice. so I don't think that is because of voice load.

Thanks man

[josef]

Hi Josef,
THere are several reasons for that problem.
- number of users
- capacity- ce iub, etc
- sgsn/ core problem
check first all ur parameter. WBTS parameter and RNC parameter.

Regards,
EJAY

All parameters are Ok. Iub capacity is ok (dual Iub with MPR 177 Mbps)

I don't think that is can be radio problem, because just some secondes before having this problem you can reach 2.7 Mbps !!!

[elkador]

Funny enough : we are experiencing exactly the same problem. I'll keep you posted. NSN RU20 though.

[sala70vic]

Throughput dropping to Zero during UE movement or in static state?
Drow the throughput against L3 events and feedback whether SCC is being performed or not.

[josef]

Funny enough : we are experiencing exactly the same problem. I'll keep you posted. NSN RU20 though.

In which country ? If there is no inconvenient for you.

Thanks.

Jo.

[josef]

Throughput dropping to Zero during UE movement or in static state?
Draw the throughput against L3 events and feedback whether SCC is being performed or not.

From my personal experience, I'm using our dongle in my hotel: so static state

What is happen is that I'm getting good throughput (enough for what I'm doing). It can be more that 2,5 Mbps and after some minutes (something like 3min) is going down to 100 Kb, to 50kb, to 4kb, to 1kb, to 0,00 kbps and It stays on 0 (HSUPA mode on dongle interface) and after some seconds it is going to umts and in umts it is starting by 0 kbps (for sure ) and it is starting to increase to 1 kbps, to 5 kbps to 10kbps to 20kbps and above one value of throughput (something like 40kbps) it is going back to HSUPA.

Without triggering from hspa to umts, throughput is staying always at 0 kbps.

I hope I explain well our problem, but it's just my experience as normal user.

Thanks guys.

[djordjebeg]

I have same problem even if i am on RU20. I have download from internet or lockal server and sudenly throughput goes to 0 kbps and stayes 6-10 seconds ( UE goeas to FACH or sometimes to cell_PCH). Than if connection to server still exists download is resumed with full speed. It happened if 2 carrier sectors are enabled and i checked 3 different modems: Huawai E182E ( cat 14), Hua ( E372 ( cat 24), and ZTE MF680 ( cat 24). radio conditions are perfect, CE , iub, radio resources usage low ( no users as network is not yet launched commercially). I opened resolving ticket to NSN and waiting for resolution. I will upstate about progress, meanwhile will try to caught internal RNC logs and IuPS GTP-U wire-shark logs and try to correlate them to get if problem is in core network side ( but i doubt).

[josef]

I have same problem even if i am on RU20. I have download from internet or lockal server and sudenly throughput goes to 0 kbps and stayes 6-10 seconds ( UE goeas to FACH or sometimes to cell_PCH). Than if connection to server still exists download is resumed with full speed. It happened if 2 carrier sectors are enabled and i checked 3 different modems: Huawai E182E ( cat 14), Hua ( E372 ( cat 24), and ZTE MF680 ( cat 24). radio conditions are perfect, CE , iub, radio resources usage low ( no users as network is not yet launched commercially). I opened resolving ticket to NSN and waiting for resolution. I will upstate about progress, meanwhile will try to caught internal RNC logs and IuPS GTP-U wire-shark logs and try to correlate them to get if problem is in core network side ( but i doubt).

Thanks man for your answer.

When you are telling that "It happened if 2 carrier sectors are enabled ", you mean that it is only on sites with 2 carriers? for us it's happen also on sites with one carrier U2100 but in all network we have another layer U900.

These are our site topologies:

- U900
- U900 + F1 U2100
- U900 + F1 U2100 + F2 U2100
- U900 + F1 U2100 + F2 U2100 + F3 U2100

Regards,

[salamanda1969]

Dear friends,

Similar events (from high throughput to zero in 5 sec) happened to me as well. Previously, my serving cell (1 carrier) could provide 500kbps to 2Mbps at night. At the end of billing month, I try to use all my credit by downloading large files/movie clips. Then can see the throughput fall off the cliff and stays at 0 kbps for long time (until I disconnect & connect again)... and this phenomenon repeats when I try downloading any large files. Any files >5MB never downloaded successfully. Very frustrating !

NSN RU20 in use.
Initially suspect QoS kicking in. But few days later, throughput back to normal again.
Now, suspect something in the core network is overloaded/dropping packets, because this issue can be repeated in other sites.
Instinct tells me RU20 needs a patch soon. :)