- This topic is empty.
24th April 2006 at 20:44 #44878pixGuest
I’m glad I found out about this forum! I’m a radio engineer working for a telecom vendor.
Does anyone have experience with dynamic Abis dimensionning, for an EDGE network ?
The dynamic Abis allows a pool of Abis timeslots to be used by any TRX which needs it. So I know the principle. But how do you know how many E1 you need ?
I guess you need to use some kind of Erlang formula ?
Thanks.26th April 2006 at 15:48 #44879PavelGuest
Transmisson capacity for EDAP can be calculated not from Erlang formula – because it is not CS traffic, but PS traffic, so you just need peak values of your traffic and prefered QoS parameters.
If you are expecting 2Mb packet traffic(EDGE/GPRS) from your BTS – you should dedicate 1xE1 for your EDAP.
Also – if you will have 2Mbit packet traffic at BTS- I can say you are happy man 🙂27th April 2006 at 13:48 #44880VanderleyGuest
When you dimention abis you should check what coding scheme do you have for example MCS9 uses 5 abis subslots. So if you have for example 4 TS for PS on every sector, for a 3sector BTS you will need just for PS 3*4*5=60subslots which is 15 timeslots on the abis.27th April 2006 at 16:28 #44881PavelGuest
2Vanderley: of course, you are right – all depends of your radio part, but I have never seen top load on all 3 cells in same time. I mean, that 15 TSL – is good, but I am not sure that it is going to be used at 100%. My prose to calculate it like 3*4*5*0.7=42 subTSL which is 10 TSL in Abis.
But if you have no transmission or PCU licensing limits- do your EDAPs as much as peak traffic can be.28th April 2006 at 06:21 #44882VanderleyGuest
If you want to show to customers that you support MCS8&9 you should be able to allocate 5 subslots per cell. Of course if you don’t want to waste resources you should use some dimensioning, but the Erlang formula is not usable in this case. We are doing the same as you wrote. Per each radio TS CS – 1 subslot; then for PS at least 5 subslots per cell regardless of the number of PS radio TS. Then if utilisation goes over 80% or you have blocking we assign more TS to that site.6th May 2006 at 01:08 #44883pixGuest
Thank you Pavel and Vanderley for the answers, sorry I didn’t check them out earlier.
Lately, I learnt that the extra Abis timeslots are NOT dedicated to one sector (like in older releases) but to the whole BTS. It’s like a pool of Abis TS. So what you need to know is, as Pavel said, the amount of data traffic expected in one BTS. For instance, at peak hour : 2 mobiles connected in MCS-9 with 4 DL PDCH in the BTS = 2 MS x 4 PDCH x 5 nibbles = 40 nibbles = 10 Abis TS.
Furthermore, the Abis nibbles dedicated to the BCCH and SDCCH can be used as extra Abis timeslots (since they are not used on the abis nibbles : they use the dedicated signalling links Abis timeslot).
With the first example :
3 sectors = 3 BCCH + 6 SDCCH = 9 empty Abis nibbles that can be used for EDGE
2 MS 4+1 = 2 x 4 radio TS = 8 Abis nibbles already reserved for voice
So in the end, you only need :
40 – 9 – 8 nibbles = 23 Abis nibbles, in other words you only need (23/4) = 6 additional Abis TS.
Instead of 12 additional Abis TS with “static abis”. Quite a great benefit ! 🙂
I’d be glad to discuss this more if anyone needs support.21st June 2006 at 13:01 #44884RoopaGuest
I would like to know that if suppose only one TS is enabled for EDGE in the air interface should the Abis also have 1 64kbps dedicated for EDGE. Do both the Air and Abis interface TS for EDGE be equal.And is there any concept of on demand EDGE allocation as in GPRS. What changes should be done in abis for that.
Thanks in advance21st June 2006 at 18:16 #44885TareqGuest
Hi Roopa. Are you working in Ericsson switch? If yes then you need to activate flexible abis in your BSC and it is introduced after R11.After that you need to change abis parameter in TG . If you dont use fexible abis then on TS will need a dedicated abis for EDGE configuration.22nd June 2006 at 05:08 #44886PixGuest
With only 1 PDCH (= 1 air TS for EDGE), you will not need additional ABIS timeslot if you activate flexible Abis. The EDGE Abis timeslot will be taken from the BCCH and SDCCH unused abis nibbles.
So no dedicated Abis TS in your case.
If you need more PDCH, then it might be needed. And you will certainly need much more PDCH than that…? Imagine 2 users sharing one PDCH… etc.23rd June 2006 at 06:23 #44887RoopaGuest
Thank you for your replies
I would like to know how many EDGE subscribers can share a single E-PDCH( both in DL/UL).
When measuring throughput of end user, is it the SNDCP or LLC or RLC throughput that gives the exact enduser throughput.
thanks in advance27th June 2006 at 14:58 #44888VanderleyGuest
The number of EDGE subscribers sharing a PDCH slot is the same as for GPRS. This is a parameter which you should set. The limitation is maximum 16MS in downlink, maximum 7 in uplink (3 bits 0-7; 0 is reserved), but the sum is max 16 as on PTCCH/D the timing advance for up to 16 MS could be transmited. Note that some vendors does not support GPRS and EDGE on the same time on one and the same timeslot.28th June 2006 at 18:03 #44889SashaGuest
May be an unrelated question regarding Alcatel Abis.
I am trying to figure out how to extract Measurement Reports from the Alcatel Abis protocol. Are there any deviations from the GSM standard implementation acording to your experience?
Thanks29th June 2006 at 06:35 #44890PavelGuest
Unfortunately, I have no expirience with Alcatel, but actually I didn’t get the point of your question.
Which deviation your are talking about? Deviation of Abis-interface, or deviation of seome measurements in Abis?29th June 2006 at 12:56 #44891SashaGuest
Abis protocol is supposed to transmit Layer 3 message called Measurement Report to assist in Handover procedure in the BSC. I have heard Alcatel does not populate this message on Abis8th July 2006 at 00:34 #44892PixGuest
The Measurements Reports MUST be transmitted on the Abis, otherwise there is no way a BSC will ever ask the mobile to make a handover.
Those measurements contain RxLev, RxQual, TA of serving cell and the RxLev of neighbors cell. Impossible to live without them 🙂
What is your current problem, actually ?