Re: yea TCP_NODELAY Re: network performance vs. linux on small transfers

1999-08-25 Thread Wes Peters
Wayne Cuddy wrote: > > Ok, if you suggested the TCP_NODELAY option you were right. Once we set this > FreeBSD sent 25 msgs/second, Linux did 22 msgs/second and HPSUX did 15 > msgs/second. (we TCP_NODELAY on all platforms) > > Is the Linux Nangle algo broken/different? It can be turned off by de

Re: yea TCP_NODELAY Re: network performance vs. linux on small transfers

1999-08-25 Thread Wes Peters
Wayne Cuddy wrote: > > Ok, if you suggested the TCP_NODELAY option you were right. Once we set this > FreeBSD sent 25 msgs/second, Linux did 22 msgs/second and HPSUX did 15 > msgs/second. (we TCP_NODELAY on all platforms) > > Is the Linux Nangle algo broken/different? It can be turned off by d

yea TCP_NODELAY Re: network performance vs. linux on small transfers

1999-08-24 Thread Wayne Cuddy
Ok, if you suggested the TCP_NODELAY option you were right. Once we set this FreeBSD sent 25 msgs/second, Linux did 22 msgs/second and HPSUX did 15 msgs/second. (we TCP_NODELAY on all platforms) Is the Linux Nangle algo broken/different? Thanks for all the help. Wayne On Tue, 24 Aug 1999, Wayn

yea TCP_NODELAY Re: network performance vs. linux on small transfers

1999-08-24 Thread Wayne Cuddy
Ok, if you suggested the TCP_NODELAY option you were right. Once we set this FreeBSD sent 25 msgs/second, Linux did 22 msgs/second and HPSUX did 15 msgs/second. (we TCP_NODELAY on all platforms) Is the Linux Nangle algo broken/different? Thanks for all the help. Wayne On Tue, 24 Aug 1999, Way