- This topic has 4 replies, 1 voice, and was last updated 11 years ago by Pix.
7th September 2009 at 15:49 #58741MacGuest
I have following queries:
1. Does the TBF drop increase as MCS increases? (assuming that radio conditions are good). If yes, why?
2. What are the reasons for TBF drop?
3. How to measure TBF establishment success rate in NOKIA system?
Thanx in advance!!8th September 2009 at 04:46 #58742BijoyGuest
1.No ideally.but please check if ur no. of re-transmissions is high for higher MCS.Also check what is your initial MCS for transmitting the first TBF.It shouldn’t be higher than MCS-5
2.Same as the reason of TCH Drop along with TBF congestion & frequent LU & Cell reselection.
3.General formula for TESR=Total successful TBF establishments/Total TBF establishment attempts.15th September 2009 at 13:11 #58743MacGuest
Retrans ARE high for higher MCS. Any issue due to that? If yes, how to reduce the same? And I am using INIT MCS = 2/1 (DL/UL). Tried using Delayed DL & extended UL TBF but to no avail 🙁
Thanks!16th September 2009 at 12:28 #58744BijoyGuest
Actually error protection bits are lesser for higher MCS,so re-transmissions are usually higher and drop rate is usually higher,this is actually a trade off between throughput & drop rate.For cells with good Um interface quality, the Link adaptation mode is usually used. For cells with bad Um interface quality, the Incremental redundancy mode is usually used.16th September 2009 at 14:26 #58745PixGuest
Just to give another solution:
Retransmissions in case of high MCS is because the C/I is just not good enough. So in those cells, you could simply limit the max MCS to MCS8 or MCS7 instead of MCS9.
The problem in EDGE is that the link adaptation is not configurable : the thresholds to increase or decrease the MCS are static.
Sometimes, I guess those tables are a little aggressive, and the MCS chosen is too high compared to the actual C/I.