How To Reduce MP Load at RNC

[sabila]

Dear all master,
please help me
my RNC has hit MPload till 90% ( used E// product 3810)
the RNC consist of ATM nodeB(80NodeB) and OverIP nodeB(120 nodeB)
i had moving 30 nodeb over ATM to other RNC but the MPLoad still hit 85%.
may be all of you can suggets me how to reduce this case :confused: :confused:

many thanks for your attention

[consultant]

Attached is SCCP MP load excell tool

[sabila]

are this tool give recomendation to solving this problem or ,tool to get MPload or other else ,please give us explanation

[consultant]

The tool is used for calculating of SCCP MP load which can help you.

[jackkarpar]

Dear all master,
please help me
my RNC has hit MPload till 90% ( used E// product 3810)
the RNC consist of ATM nodeB(80NodeB) and OverIP nodeB(120 nodeB)
i had moving 30 nodeb over ATM to other RNC but the MPLoad still hit 85%.
may be all of you can suggets me how to reduce this case :confused: :confused:

many thanks for your attention

Oh! i think we meet the same problem, do you have low rrc setup rate in your network when the mpload was high?

[jackkarpar]

The tool is used for calculating of SCCP MP load which can help you.

Since that he have known the mpload was high, i don't think he need your mpload calculation tool. But that would helpful for others. Thanks all the same! And Would you share your experience in solving high mpload problem?

[sabila]

Oh! i think we meet the same problem, do you have low rrc setup rate in your network when the mpload was high?

yes degradation of rrc

[RFDude_PMA]

Well mate..

At last somebody with the same e/// problem.. Well In my network Mpload is due to signalling coming from switching or rrc connections. When we program URA in our network we manage to reduce the mpload. But at the end is RNC capacity the all problem.

We have to split the load.. We add extension modules. But the most suitable way is to reduce the switching from rab to rab. For this you can verify the timers and buffers.

BRGS

[jackkarpar]

Well mate..

At last somebody with the same e/// problem.. Well In my network Mpload is due to signalling coming from switching or rrc connections. When we program URA in our network we manage to reduce the mpload. But at the end is RNC capacity the all problem.

We have to split the load.. We add extension modules. But the most suitable way is to reduce the switching from rab to rab. For this you can verify the timers and buffers.

BRGS

hi, RFDude_PMA!
what do you mean by programming URA? What parameters should be adjusted? Could you please give us more detail about that? Thank you.

[RFDude_PMA]

Two parts here:

1. URA activation is made at RNC level activating uerc 21
2. Then every Node need to be assign a ura_ref that initially can be the same as rac
3. At the RNC level are 2 timers that need to be change
a. inactivitytimer 10
b. inactivitytimerpch 30

AT the RNC border is recomended to create a ura too
brgds

[RFDude_PMA]

Hi Jackkarpar

Something to keep in mind that URA activation will change drop connections formula!! You have P7?

___________________________________________
If you find this information useful pls add some reps

[jackkarpar]

Hi Jackkarpar

Something to keep in mind that URA activation will change drop connections formula!! You have P7?

___________________________________________
If you find this information useful pls add some reps

hi, we have upgraded to w10b, but we found that the rlc buffer's size was too low in our network lead to high frequency channel switches between fach and dch. the mpload drop to 60% from 85% after enlarge the buffer. Thanks!

[sabila]

yes same I had problem with capacity ,there i had new RNC 3820,but still much problem at this new RNC.

we had upgrade to w10 too,but no improvement,

BTW how to see the RLC buffer at rnc and how to set it
and how to calculate how big we should set it

may u can give me explaination

many thanks for yout help

[RFDude_PMA]

Well Thanks mate..

We had the same bad accesibility problems due to mpload rejections.. We had only one RNC with 100 nodes 3sec each. Then the fast solution was URA which helps for a while, but the traffic increase so the next step was new RNC.. With a new RNC 3820 with 2 extension racks it solved it but, it was just temporally. The traffic increase and then the two rnc one with 139 and the other 80 have the same mpload rejections problems.. The buffers are double the recommend values..

What is on the table as urgent is the change to W10 because according to E// specs it have more capacity.. JAJA the same thing was with 3820.. but we try everything..

[sabila]

so what parameter should change
are this parameters?to be double from rcomend value

dlRlcBufUpswitch
ulRlcBufUpswitch
dlRlcBufUpswitchMrab
ulRlcBufUpswitchMrab

many thanks

[RFDude_PMA]

What we have is this..

dlRlcBufUpswitch 1000
dlRlcBufUpswitchMrab 5
ulRlcBufUpswitch 1024
ulRlcBufUpswitchMrab 256

How much is your rab traffic? and how many nodes your rnc have?

[sabila]

overall 16.000.000 erlang
with 200 nodeB

[RFDude_PMA]

hi Sabila..

Do you have 16millions of rab erls and 200 nodes in one RNC?

[jackkarpar]

hi Sabila..

Do you have 16millions of rab erls and 200 nodes in one RNC?

can't believe it. we only got 6~7 hundred erlang per rnc:( in bh

[sabila]

yes
this data i got from NPI e*******

[RFDude_PMA]

ok.. Well we have less that 4k Rab erls on one RNC 3810 and we have 85 nodes 3815 connected to it. The issue is that for us this level of traffic over capacity is causing mpload rejections that reduce the accesibility.

If you have 16M of Rab Erl on one RNC well share with us the channel switchin g parameter values..

[Nyquist]

Hi Sabila.
Once I heard about a similar problem with E///3810 with ATM and IP nodesB. That time was due to a lack of IP host definition in all subracks even if do not have any IP interface board. As you say, they moved ATM nodes to another RNC but load still remains.
Sorry but I do not have more details.
I hope this could help you.
Nyquist
:D