In message <87k6fzww0w.fsf@snark.piermont.com>, "Perry E. Metzger" writes:
>
>"Steven M. Bellovin" <smb@cs.columbia.edu> writes:
>>>Never mind the recovery time difference between DOWN and UP, look at
>>>the pattern where they go down.  They are remarkably aligned, 3 hours
>>>and 4-5 seconds apart. =20
>>
>>>Figure out what happens at that interval, and you have your cause.
>>
>> What is the dhcp lease rebind time?
>
>I hard code the address on that network.

Hmm -- anything going on on the server end?
>
>BTW, I tried using the device a bit today at Columbia University --
>very bad performance. The device keeps pathologically scanning from
>one modulation mode to another very rapidly.
>
And I had no problem on campus today -- I couldn't get a single twitch 
from 80211watch; it locked in immediately and stayed there.

		--Steven M. Bellovin, http://www.cs.columbia.edu/~smb