Control Slots - the magic number?

I’ve read the manual, and see where it kind of labels control slots and how many you should have.

Has anyone verified this? Or is there a different "magic number"

ie:

1 control slot = 1-50 customers
2 control slot = 50-100, etc.

What have you seen?

It’s true. We had 1 set on cluster with 80-90 SMs per sector. We had many reregs. When we put 2 most of the problems disaperd.

we are currently at 4, adding 1 at a time to get the most stable performance. I was told that one wisp found the sweet spot at 6. i was told to use the FECuToss to judge how many you need.

Didn’t I read something somewhere about control slots not being as important in HW scheduling?

And what about the FEC utoss?? Just found it in the AP Bridge stats. Am I looking at the right thing? And assuming that I am, what does it mean?

Thanks!

They are very important. they handle bandwidth requesta and grants to the SM’s.

FECuToss means the AP dropped packets due to the fact it hould not handle the amount of unicast (u=unicast b=broadcast) data comming into it. you will probably always have some, the key is to minimize them.

So… what’s an acceptable rate? I have an AP with about 100 clients on it, and just bumped from 2 control slots to 3. Uptime is 20 hours 12 minutes, and FEC utoss is now at 1192. Is that a lot, or is that acceptable?

Thanks for the info!

If it helps any I have 110 customers on one Ap we run 5 control slots

I add up the total FECuIn and FECuOut and divied by FECuToss and try to feep it under 1%.

Thanks, Mike, that gives me something to go on!