E******* RNC MP Load component analysis

[electron]

Hi guys,

Could some body explain how to understand/calculat MP load component model for E******* RNC . I mean how to calculate contribution for each different signalling , channel switching , different RAB establishment,... contribution to increase MP Load to current existing level?


Thanks

[babak1349]

hello Electron,
To calculate the MP load based on events , you need to know how much each event (RRC signaling, SHO, Channel switching, etc..) adding to GPB load So by knowing those values you can estimate expected GPB( Mmp load)
You can look at Alex inthe section (Dimensioning Description RNC 3820, W12) and you can see better description
But what is the benefit of such calculation ? dimensioning your RNC boards ? You need to know exact number of events to do so Rough estmiation can't help much. If you have existing RNC it is better too look at exact load generate by traffic
Furthermore you need to look at the Load management of GPB/ Modules as well to help you to understand better the mechanism even though it is not very detailed in Alex


Cheers

Hi guys,

Could some body explain how to understand/calculat MP load component model for E******* RNC . I mean how to calculate contribution for each different signalling , channel switching , different RAB establishment,... contribution to increase MP Load to current existing level?


Thanks

[electron]

Salam Babak,


Thanks for your information . I actually gonna analysis RNC MP load component to get what action(RAB Est,Channel switching,..) is the most contributor for MP load rise. Regarding Di mentioning i have been looking to it long time ago but actually it will not help in this regard( I actually gonna chart contribution of each element for MP load hourly)

hello Electron,
To calculate the MP load based on events , you need to know how much each event (RRC signaling, SHO, Channel switching, etc..) adding to GPB load So by knowing those values you can estimate expected GPB( Mmp load)
You can look at Alex inthe section (Dimensioning Description RNC 3820, W12) and you can see better description
But what is the benefit of such calculation ? dimensioning your RNC boards ? You need to know exact number of events to do so Rough estmiation can't help much. If you have existing RNC it is better too look at exact load generate by traffic
Furthermore you need to look at the Load management of GPB/ Modules as well to help you to understand better the mechanism even though it is not very detailed in Alex


Cheers

[babak1349]

Salam,
Yes.. It is all dependent on how you are doing parameter setting For example how is your SHO or Channel switching setting... Of course RNC load is coming in lower priority due to over dimensioning normally and you set the parameters to maintain the best NW quality.
The point is what is going to be the outcome of your efforts ? If SHO is the most contributor in Mmp load ? what you are going to do about it ? Reducing SHO may cause more drop call.
Please also remember that the load will be shared evenly among of RNC modules and that's why you see even load between different Mmp ( I think there is also 10% margin). Ofcourse C1 and C2 GPB boards will have different load due to type of tasks they do


Cheers

[electron]

Actually regarding resource sharing between modules i am aware about it but main intention of mine is to have clear picture for MP load increment factor as well as their contribution for different RNC anyway i will find the solution lately . Thanks for your information

Salam,
Yes.. It is all dependent on how you are doing parameter setting For example how is your SHO or Channel switching setting... Of course RNC load is coming in lower priority due to over dimensioning normally and you set the parameters to maintain the best NW quality.
The point is what is going to be the outcome of your efforts ? If SHO is the most contributor in Mmp load ? what you are going to do about it ? Reducing SHO may cause more drop call.
Please also remember that the load will be shared evenly among of RNC modules and that's why you see even load between different Mmp ( I think there is also 10% margin). Ofcourse C1 and C2 GPB boards will have different load due to type of tasks they do


Cheers

[ivanbrij]

Dear Electron,

In the link below you there is a good document about that. Please add REP.

http://www.4shared.com/office/J-NGrq...imization.html

BR,

Ivan

[elpirata]

Thanks a lot for this doc.

And perhaps do you have any exemlpes how to use this pm counters?
I need to check IuCS interface.

P.S.
Pehaps, like this:
pmx IuLink pmInFrames(Success)*$ -m 2.5

A?


Thanks.