marcengo
2016-03-03, 02:53 AM
Hi there, ninjas.
I`m currently trialing the feature LTE908 ANR Inter-RAT UTRAN - Fully UE-based in NOKIA RL70. This feature is based in active UE measurement and reports of UMTS cells (reportCGI).
Issue: Drop after the feature activation.
Description:
The trials are being done like this:
1) Delete IRAT NL.
2) Activate the feature with minNotActivatedUtraRSCFS=0 (periodicity the eNodeB command UEs to measure, meaning, always measure) for three hours to speed up NL recreation.
3) Change minNotActivatedUtraRSCFS=4h.
4) A couple days after disable ANR IRAT.
The feature itself presents good results in terms of neighbor list creation (quality wise), but it presents intense Drop degradation while minNotActivatedUtraRSCFS=0 and some Drop degradation with minNotActivatedUtraRSCFS=4h. The results show a clear variation according to minNotActivatedUtraRSCFS=0 than 4h. Later on when ANR is disable, Drop return to equal or lower levels.
The initial suspect was regarding ANR x DRX somehow not being aligned (not enough time to measure UMTS while on DRX Sleep) which could cause UEs to Drop. For that matter, before changing the DRX Cycle configuration, we did the following trial:
1) Delete IRAT NL.
2) Activate the feature with minNotActivatedUtraRSCFS=0 (to stress the issue of Drop)
3) Three hours later disable DRX
This time the same increase was seen with minNotActivatedUtraRSCFS=0 and later than when DRX was disabled, improvement in the Drop, but once again, not in the same levels as before the ANR was activated.
In this sense my conclusion is: DRX affects but it is not the only factor affecting the Drop.
Have you ever seen similar behavior in other NOKIA networks? What about similar features in other vendors?
Unfortunetelly at this time only RRC-level logs are available upon request.
I`m currently trialing the feature LTE908 ANR Inter-RAT UTRAN - Fully UE-based in NOKIA RL70. This feature is based in active UE measurement and reports of UMTS cells (reportCGI).
Issue: Drop after the feature activation.
Description:
The trials are being done like this:
1) Delete IRAT NL.
2) Activate the feature with minNotActivatedUtraRSCFS=0 (periodicity the eNodeB command UEs to measure, meaning, always measure) for three hours to speed up NL recreation.
3) Change minNotActivatedUtraRSCFS=4h.
4) A couple days after disable ANR IRAT.
The feature itself presents good results in terms of neighbor list creation (quality wise), but it presents intense Drop degradation while minNotActivatedUtraRSCFS=0 and some Drop degradation with minNotActivatedUtraRSCFS=4h. The results show a clear variation according to minNotActivatedUtraRSCFS=0 than 4h. Later on when ANR is disable, Drop return to equal or lower levels.
The initial suspect was regarding ANR x DRX somehow not being aligned (not enough time to measure UMTS while on DRX Sleep) which could cause UEs to Drop. For that matter, before changing the DRX Cycle configuration, we did the following trial:
1) Delete IRAT NL.
2) Activate the feature with minNotActivatedUtraRSCFS=0 (to stress the issue of Drop)
3) Three hours later disable DRX
This time the same increase was seen with minNotActivatedUtraRSCFS=0 and later than when DRX was disabled, improvement in the Drop, but once again, not in the same levels as before the ANR was activated.
In this sense my conclusion is: DRX affects but it is not the only factor affecting the Drop.
Have you ever seen similar behavior in other NOKIA networks? What about similar features in other vendors?
Unfortunetelly at this time only RRC-level logs are available upon request.