We believe this is truly an outside offender in 2.4 GHz. I have personally seen a carrier that is several times more power than anything I have ever seen. I only saw it for a brief instant though. This interference just does not last long enough to be caught. The high latency is caused by retransmits but I am sure outside interference is what is leading to the need for frames to be sent again. This effects all channels across the 2.4 GHz bands. We have seen the noise floor jump up higher than our radio power levels when this problem happens. What ever is causing this is running higher power than anything we have in the field. We will look at anything, though, to help troubleshoot and I appreciate your ideas.

Mac Dearman wrote:


I tend to believe you will find your answer on your network -vs- "big bad
leak" somewhere and the only real suggestion I can offer you would be to do
what we do here when we start having weird issues - - and that is to be
ready to start unplugging Back Hauls (one at a time) out of their
switch..Etc when the trouble starts. Have a ping running to watch for the
latency to disappear when you unplug the offending Back Haul. Once you
narrow it down to the right leg of the network you can reboot one AP at a
time until you find which AP it is coming from by running an extended ping
on that leg of the network

Don't pull a "big dummy" like we did here for weeks!! Unplug the servers
one at a time too!!! That is where we found one of the servers we host was
our culprit. There traffic was of such that it was flooding our switches - -
(dirty suckers) :-)

GL - & keep us informed as to what you find

Mac Dearman

-----Original Message-----
Behalf Of David E. Smith
Sent: Friday, October 27, 2006 4:10 PM
To: WISPA General List
Subject: Re: [WISPA] "The Gremlin," redux

On Fri, October 27, 2006 3:11 pm, Eric Merkel wrote:

1) Turning off inter-BSS Relay

Already done, on most towers. (We do have a couple of towers where one
business, with two locations, wants to do VPN-type stuff between 'em.)

2) We block all the typical MS ports(135-139) which broadcast all the
time via iptables

Done. We block 135-139, 445, and a couple other ports, both TCP and UDP.

3) Packet shape all connections via CBQ on the AP itself to limit how
much bandwidth any one customer can consume

Mostly done. (For historical reasons, some of our customers are still part
of a giant bridged network, and their traffic is shaped in our office not
at the AP, but those customers are relatively few and growing fewer by the

David Smith
WISPA Wireless List:



Reply via email to