Return-Path: Received: by vger.rutgers.edu id <154534-8316>; Fri, 11 Sep 1998 18:21:31 -0400 Received: from mail.cyberus.ca ([209.195.95.1]:55562 "EHLO cyberus.ca" ident: "NO-IDENT-SERVICE[2]") by vger.rutgers.edu with ESMTP id <154629-8316>; Fri, 11 Sep 1998 17:49:59 -0400 Date: Fri, 11 Sep 1998 20:39:46 -0400 (EDT) From: jamal To: linux-kernel@vger.rutgers.edu Subject: Re: my broken TCP is faster on broken networks Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-kernel@vger.rutgers.edu Content-Length: 981 Lines: 30 Theodore Y. Ts'o (tytso@MIT.EDU) Fri, 11 Sep 1998 18:36:43 -0400 >I don't normally follow the TCP implementor's working group (*), so I >don't know if they actually followed through on his suggestion; >Matthais's note seems to indicate that they did. I think it is just considered good practise to punish misbehaving users at the moment. Not sure if any vendor has implemented anything yet. There is an informational RFC out: http://info.internet.isi.edu:80/in-notes/rfc/files/rfc2309.txt Unfortunately, they dont mention remedies for misbehaving flows. They do suggest RED; however, there are variants of RED that exist with specific intention of punishing misbehaving users. A really interesting pointer page is at: http://www-nrg.ee.lbl.gov/floyd/tcp_unfriendly.html cheers, jamal - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/faq.html