PHWinfo - Afficher un message

01/11/2006, 21h33

Krit <krit_kasemosoth@hotmail.com> wrote:
> I already used TCP_NODELAY option to disable Nagle's algorithm in
> TCP layer, however, the problem still happen; the server did not
> send the next timestamps to the client until it receive TCP ACK from
> the client.

Did you set it at _both_ ends?

Apart from the suggestion to use UDP (and the implicit "provide your
own retransmissions") you might make sure that something like a
botched Appropriate Byte Counting congestion control isn't in effect.
Grep for "tcp" in sysctl -a and see if there is anything along those
lines.

Thankfully, netperf TCP_RR doesn't rely on one-way latency measures

rick jones
--
Wisdom Teeth are impacted, people are affected by the effects of events.
these opinions are mine, all mine; HP might not want them anyway...

feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...

01/11/2006, 21h33	#3
Rick Jones Messages: n/a Hébergeur:	Re: delayed ACK in TCP problem Krit <krit_kasemosoth@hotmail.com> wrote: > I already used TCP_NODELAY option to disable Nagle's algorithm in > TCP layer, however, the problem still happen; the server did not > send the next timestamps to the client until it receive TCP ACK from > the client. Did you set it at _both_ ends? Apart from the suggestion to use UDP (and the implicit "provide your own retransmissions") you might make sure that something like a botched Appropriate Byte Counting congestion control isn't in effect. Grep for "tcp" in sysctl -a and see if there is anything along those lines. Thankfully, netperf TCP_RR doesn't rely on one-way latency measures rick jones -- Wisdom Teeth are impacted, people are affected by the effects of events. these opinions are mine, all mine; HP might not want them anyway... feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...