I fact, I don't know if something has changed in the tcp implementation of
2.0.33 kernel, but it seems to me like the tcp rtt timers are not very
much smoothed, like if after 2 "fast" acks it goes down very quickly, like
if the variance of the rtt where not take in count.
I'm looking tcp_timer.c in linux source, but I am not familiarized with
it. Could somebody write me/us if this behaviour of the tcp is right ?
Another thing, maybe in relation. I set the irtt with the route command.
Maybe the problem is here, he have seen that this program has some
bugs/not well documented (for example the "mss" wich in fact is the mtu).
/sbin/route add -net 44.133.228.0 netmask 255.255.255.192 window 4320 mss
512 irtt 60000 dev bc0
Thanks !!
Julián
-.-