This document discusses and highlights the known issues related to SDCCH drop rate customisation in NSN network for Bakcell. The changes required in existing formula for Bakcell network in comparison to exSiemens formula are also proposed. ecommendations and proposal for comparison are also pro!ided.
setting
a
threshold !alue
for
2.Kno$n I%%!&% R&'(t& to S)&"#*#" Co!nt&r% The " counters discussed #elow ha!e high !alues which do not re$ect the actual situation in the network. The reasons and %ndings from ex-Nokia NSNs&stems are also pro!ided #elow against each counter. 1. SDCCH Radio Fail: The counter is pegged for causes where the 'S user disconnects a call attempt during the SDCCH phase. (lease see description #elow. Hence this counter does not necessaril& re$ect all failures due to radio en!ironment. )ollowing is the detailed explanation of the counter* Counter +D * , Counter Name * SDCCH/0D+1/)0+2 Description* 0 co!erage pro#lem3 for example 4 'S mo!es out from timing ad!ance Common in connection with co!erage pro#lems. Triggered also if the 'S user clears the call in the SDCCH phase.
2. SDCCH Abis Fail: The counter is also pegged for causes other than actual drops on SDCCH due to 0#is failure. 5host sei6ures ma& ha!e a #ig impact. The s&stem acti!ates multiple SDCCH channels against multiple requests for channel
percentage settled down to around ,?-" per cent. +n BTS B@3 BTS 0CH detection was impro!ed3 and %gures well under , per cent are now t&pical.
SDCCH d"o# $% S& 'sd"1a( =se* To follow up the performance of SDCCH from a technical point of !iew. NSN 9xperiences on use* ,; High SDCCH drop rates usuall& result from ghost accesses. 0 BTS decodes them from en!ironmental or #ackground noise and %lters out most of them. Howe!er3 all of them cannot #e %ltered out and the 0CH request is passed on to the BSC for processing and for the allocation of a SDCCH channel. The counter ghost_ccch_res :; is updated each time a channel required is reAected #ecause of an in!alid esta#lishment cause. +n 5S' ph., there exist altogether eight esta#lishment causes3 three of which are unde%ned as in!alid3 for example3resulting in that this counter shows onl& ? of all the ghost accesses the BTS has decoded. )or the rest3 a SDCCH is allocated and this will result in sdcch_abis_fail_call failure. Because of ghost attempts the SDCCH drop ratio is high with low tra>c. 0s the amount of call attempts increases3 the in$uence of ghosts #ecomes smaller and the drop ratio approaches its real !alue. "; The rate of ghosts coming to SDCCH dropped when BTS B@ with impro!ed ghost %ltering was taken into use. 0lso one important thing +n SDCCH failure counters it is not possi#le to separate 2= and call sei6ures.
situation in the (+Es. 'oreo!er it is also not included in the ex-Siemens formula for SDCCH drop rate. During formula mapping also it was remo!ed from the Nokia formula to #e a#le to make similar to Siemens formula. The ex-Siemens formula is gi!en #elow* S)m ''*RCLRR+,SD-ASSFAILRFOCSD(/*S0CCHPC where NC29F is the GClear equest message sent from BSC to 'SC on the 0-interface to clear the dedicated channel due to an& failure on the BSS part. 0n& failures due to 0-interface are not included according to this formula. NS=CCH(C is the num#er of immediate assignments for signalling channels. )or ex-Nokia NSN formula3 the SDCCH drop rate for BSS part onl& should exclude 0-interface failures similar to ex-Siemens. Therefore onl& the following counters are recommended to #e part of the formula. This was also the formula originall& proposed as a matched formula against ex-Siemens one. sum(a.sdcc"(radio()ai + a.sdcc"(r)(od("o + a.sdcc"(*ser(act + a.sdcc"(bcs*(reset + a.sdcc"(netw(act + a.sdcc"(abis()ai(ca + a.sdcc"(abis()ai(od + a.sdcc"(bts()ai + a.sdcc"(apd()ai - b. +%,0,(-/R-D ) 100 * -------------------------------sum(sdcc"(assi!n +sdcc"("o(sei1 )
5.R&"o--&n& N&$ For-!'( The following formula eliminates the maximum possi#le duplications that are pegged in the counters and gi!e the closest possi#le !alues to re$ect the actual situation in the network for ex-Nokia NSN BSS. Still it should #e noted that it is not ,I correct as it is not possi#le to eliminate all the false pegging of counters.
Jhere C is the part 0#is/)ail/call which occurs #efore esta#lishment indication* C K sdcch/assign L sum of all sei6ures K sdcch/assign L :succ/sei6/term M succ/sei6/orig M sdcch/call/re/est M sdcch/loc/upd M imsi/detach/sdcch M sdcch/emerg/call;
6. R&%!'t% (n Pro)o%('%
The issues related to wrong pegging of !arious counters is highlighted and information is pro!ided as per ex-Nokia NSN research and technical %ndings.
The exclusion of 0-interface failures is necessar& as explained in detail alread&. +t should #e noted that it was not part of the original formula proposed for Bakcell network that was mapped to #e according to exSiemens formula. NSN 9x-Nokia also does not recommend to include it for SDCCH drops at BSS part.
The new recommended formula gi!es the closest and most exact possi#le !alues from ex-Nokia NSN BSS s&stem.
There will #e a dierence in the (+ computed as compared with ex-
Siemens as the pegging of these counters and segregation is !er& dierent in ex-Nokia NSN. 5enerall& it is expected to #e much higher in ex-Nokia as compared with ex-Siemens BSS. +t should #e noted that for other networks using ex-Nokia equipment3 it is generall& higher #ut this does not re$ect the actual on ground situation.
9!en with the new formula3 the !alues will not #e in range of !alues o#tained #& ex-Siemens BSS and hence a threshold !alue is set to #e met as per acceptance process.
The !alues actual network performance can also #e Audged #& considering +mmediate 0ssignment Succ ate and 0ssignment Succ ate.
The recommended threshold !alue for SDCCH drop rate is 2.$. This should #e de%ned as the acceptance threshold and its aect in CSS should also #e accommodated for.