Re: [Madwifi-devel] IBSS BSSID fix for large networks - 400 nodes

[Georg L. <ge...@bo...>]

Hello,

* Derek Smithies <de...@in...> [2007-07-06 08:30]:
> We considered the appropriate values for the rate control algorithm. How 
> frequently should it sample other possible rate values. This particular 
> value is seriously important.
> If it samples other (and probably non optimal values) rates too 
>   frequently, the overall throughput will drop.
> If it does not sample other (and possibly non optimal values) frequently 
>  enough, the algorithm will not respond to changes in the environment.

Exponential sampling intervals come to mind as a strategy... i.e. double
the time until the next sampling every time when the current rate is the
best one.

...which of course wouldn't fix the suddenly escaping notebook user
scenario you described.

IIRC the hardware contains provisions to reduce the rate for packet
retries, and you could react really fast when you detect a packet drop.
This should help resolving the TCP congestion mis-detection.

kind regards,

Georg
-- 
|| http://op-co.de ++  GCS/CM d? s: a-- C+++ UL+++ !P L+++ E--- W++  ++
|| gpg: 0x962FD2DE ||  N++ o? K- w---() O M V? PS+ PE-- Y+ PGP++ t*  ||
|| Ge0rG: euIRCnet ||  5 X+ R tv b+(+++) DI+(+++) D+ G e* h! r* !y+  ||
++ IRCnet OFTC OPN ||________________________________________________||