Generic selectors
Exact matches only
Search in title
Search in content
Search in posts
Search in pages

Reply To: High SDCCH Assignment Failure

#53417
SA
Guest

Hello to All.

I’ve been working with Ericsson GSM Networks for some time and since mid 2007 we noticed that we were expecting a very similar problem, or maybe the same one. At least in 3 countries I could certainly say that the problem is present.

The diagnostic is the following:

– Raise over the normal values of the following counters:

CNROCNT: Number of accepted random accesses
RAOTHER: Number of random accesses, Location Updates
CCALLS: Channel allocation attempt counter (on SDCCH)
CSIMMASS: Number of CS IMMEDIATE ASSIGNMENTS sent on the CCCH
CESTIMMASS: Number of SDCCH establishment failure due to timeout after sending Immediate Assignment, timer T3101 expired

– Fall of the SDCCH Access (SDCCH Assignment)
In Ericsson the formula is (CMSESTAB/CCALLS) x 100%
It falls because the number of Attempts on SDCCH (CCALLS) grows without control, but the number of successful attempts on SDCCH (CMSESTAB) remains the same.

– With ABIS traces, Ericsson found that (just like Flo said) the ms is not authenticated yet so there is no way to find the IMEI or the phone model. Ericsson agrees that it is a specific model or models from our network or from the competition. However A VERY INTERESTING FACT is that all the attempts come from a specific Timing Advance (TA) giving strength to the theory that it is indeed a mobile or group of mobiles that produce this effect. Also, if you use a MRR (Measurement Result Recording) or something similar for a specific hour, if the cell has the problem present, a certain Timing Advance grows beyond all others.

– The problem may be present for 1 or several hours.

– If the problem is caused by Location Updates, then one or more than one of the following procedures is “guilty”: normal registration (with the change of a Location Area), periodic registration and IMSI attach. We moved the Timer for the Periodic Registration from 1 hour to 4 hours in a very large group of cells and the problem persisted. Also we turned off the IMSI Attach procedure in a complete BSC for 1 day and the problem endured. And we have seen the problem in areas where there is no border of a location area. As you can see, it is a very complicated issue!

– It happens every day in at least 5% of the sectors of the network.

¿Any thoughts? Regards,

SA