Date: Sun, 30 May 1999 13:31:40 -0400 (EDT) From: David Gilbert <dgilbert@velocet.ca> To: Luigi Rizzo <luigi@labinfo.iet.unipi.it> Cc: dgilbert@velocet.ca (David Gilbert), wollman@khavrinen.lcs.mit.edu, freebsd-net@FreeBSD.ORG Subject: Re: Linux and Solaris tcp beating FreeBSD 4:1 Message-ID: <14161.30204.750053.989577@trooper.velocet.ca> In-Reply-To: <199905291654.SAA15653@labinfo.iet.unipi.it> References: <14160.1181.137755.84641@trooper.velocet.ca> <199905291654.SAA15653@labinfo.iet.unipi.it>
next in thread | previous in thread | raw e-mail | index | archive | help
>>>>> "Luigi" == Luigi Rizzo <luigi@labinfo.iet.unipi.it> writes: >> Since the real solution for this is some time in the future, what >> can I hack on quickly? If I can't change the timeout speed, can I >> change some other parameter that will help here? I'm taking a lot >> of flack that netcat port forwarding on a Sparc 1+ can beat our >> FreeBSD K6-2/400 Luigi> one wonders what is your config and why you have high losses Luigi> (congestion or what ?). Lowering the MTU in some cases might Luigi> help because it keeps more packets in flight and lets fast Luigi> retransmit act earlier. As I explained in my first message, we're providing our selves with ADSL. Since we are further than we should be from the other end, we only barely get a connection. However, this gives us 816K up and 680K down for about $10/month. It is worth some optimization on the link. Fast retransmit is an something that reacts in less than a second? Dave. -- ============================================================================ |David Gilbert, Velocet Communications. | Two things can only be | |Mail: dgilbert@velocet.net | equal if and only if they | |http://www.velocet.net/~dgilbert | are precisely opposite. | =========================================================GLO================ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?14161.30204.750053.989577>