Generic selectors
Exact matches only
Search in title
Search in content

Reply To: MC02d – ALU, help, please!


Hello, Pix!

I’m glad to cheer you up :)))

But I think you be laughing even more when you know the whole story. Where should I start?

Maybe here – in our company in radio subsystem we use two vendors equipment: alcatel and huawei.
And huawei is very aggressive, there are lots of its fans unfortunately. One of the argument in favor of huawei that it has better performance – i.e. better KPIs. Some of my colleagues had done the great job they figured out what was the thing. The problem was in methods of calculation it’s expected but they proved it. They also offered some ways to correct formulae – in order to have possibility compare comparable things. But nobody cares, all our standards are sent from headquarter and are like Bulls of Pope.
But digging in counters and all that stuff they also found the other curious thing. There are many cells where CSSR is above 100% because Imm_Ass_Succ_Rate is above 100%.

The formula we use Imm_ASS = (MC148 – MC02a)/ (MC01 + MC02 – MC964)

And our guys found out that imm_Ass > 100% when MC02d is present and significant in the cell.
After your explanation I understood at last what it means :))
So increasing the MC02d will help to improve our kpis… that’s the matter

Now – about really interesting stuff. I chose for my *research* 4 cells where mc02d is present permanently (i.e. every hour) and significant. Let’s take one of 4 cells – I studied it very well 🙂
So we have:
– mc02d is about 120-130 events per hour . even in the night!
– At the same time the number of TCH normal assignments is much less: 30-40 per hour during the daytime and 3-7 in the night
As you said *follow-on* means that MS initiating a call right after LU. So I think we should see that the number of TCH assignments must be equal or bigger than mc02d
Did I understand you right?
But it doesn’t happen. It seems that some MS (hm-m-m, not- some devices ) that is located at LAC boundary (for example) is performing LU very very often (mc02d per hour). But they are made in china and put bit FOR for *1* in CHNREQ message, not *0* like all other normal MS – it’s not my thought 🙂
I’ve just citate my colleague 

A GSM bedtime story :))))