]) by chopin.merit.edu (MOS 3.8.2-GA) with ESMTP id AFJ81506; Thu, 12 Jul 2007 17:14:45 -0400 (EDT) Received: (from steve@localhost) by mrserver.telecomplete.net (8.13.8/8.13.8/Submit) id l6CLEgAS014617; Thu, 12 Jul 2007 22:14:42 +0100 Date: Thu, 12 Jul 2007 22:14:42 +0100 From: Stephen Wilcox <steve.wilcox@packetrade.com> To: Iljitsch van Beijnum <iljitsch@muada.com> Cc: Philip Lavine <source_route@yahoo.com>, nanog <nanog@merit.edu> Subject: Re: TCP congestion Message-ID: <20070712211442.GC14059@MrServer.telecomplete.net> References: <796919.42253.qm@web30811.mail.mud.yahoo.com> <922B68D8-BCB4-40B8-88 5A-2459BDFEDA34@muada.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <922B68D8-BCB4-40B8-885A-2459BDFEDA34@muada.com> User-Agent: Mutt/1.4.2.2i Sender: owner-nanog@merit.edu Precedence: bulk Errors-To: owner-nanog@merit.edu X-Loop: nanog X-Junkmail-Status: score=10/50, host=mozart.merit.edu X-Junkmail-SD-Raw: score=unknown, refid=str=0001.0A090204.46969D83.005B:SCGAP167720,ss=1,fgs=0, ip=198.108.1.26, so=2006-09-22 03:48:54, dmn=5.3.14/2007-05-31 Status: RO X-Status: X-Keywords: X-UID: 26 On Thu, Jul 12, 2007 at 10:54:56PM +0200, Iljitsch van Beijnum wrote:

On 12-jul-2007, at 22:27, Philip Lavine wrote:

...

I just don't understand how if there is 1 segment that gets lost how this could translate to such a catastrophic long period of slow- start. How can I minimize the impact of the inevitable segment loss/out of order over a WAN. Is QoS the only option?

Also note that minimal amounts of cell loss on ATM create huge amounts of packet loss at the IP layer.

from the phrases used ('catastrophic') my feeling is that perhaps Philip isnt un derstanding that on a high speed TCP transfer a single missing bit of data can c ause the whole thing to stop and restart transmission from slow Steve