- This topic is empty.
23rd October 2008 at 13:46 #54292Attempts and TBF_CongestionGuest
Is there any body to know…In my network there is some cells with high values of tbf_attempts and some with high value of tbf-congestion rate_ul. What is the reasons for high value of attemps(7000-12000) per hour and high values of TBF_Congestion rate in uplink?
Thank you in advance23rd October 2008 at 15:23 #54293pixGuest
high value of attempts in UL, because usually the attempt, if it is failed, will be repeated automatically by the MS, several times.
Congestion of the UL radio resources –> more attempts in UL.
I think if you fix the congestion, the number of UL attempts will become more reasonnable. However, normally, you should see DL congestion rather than UL congestion. Because the DL channels are always more loaded than the UL channels, in “packet transfer” world.
Can you isolate the cause of the congestion ? Radio congestion ? Abis congestion ? GPU congestion ? Gb congestion ?24th October 2008 at 07:17 #54294abbaGuest
Thank you PIX …
This congestion are PDT congestion, but I think that there is enough capacity in PCU boards.It is strange that the congestion are generated from cells with low data traffic..May be there are some parameters for resource distribution that we should change?24th October 2008 at 13:49 #54295PixGuest
What is PDT ?
Low “useful” data traffic is one thing, but there might be a high amount of signalling data. What is the amount of RLC blocks transmitted for these cells, compared to other neighbor cells ? Is it less ?
Are you saying that within the same area (same BSC, same PCU), some cells are congested, and others are not, while the GPRS traffic (signalling + data) is approx. the same ?
In this case, it’s probably, either:
different GPRS parameters (MCS adaptation, PDCH/users, TBF/PDCH, etc.)
wrong load sharing of the PCU (is it something you configure ? in alcatel, that’s automatic)
or perhaps it’s just that the TCH (voice) traffic is just using up all radio resources, therefore the GPRS looks congested.
There are so many possible problems there… you really need to dig deeper in your QoS indicators 🙂