From owner-freebsd-hackers@FreeBSD.ORG Thu Jul 10 12:43:16 2003 Return-Path: Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 499ED37B401 for ; Thu, 10 Jul 2003 12:43:16 -0700 (PDT) Received: from Chow.corp.media.net (rottie.media.net [66.113.65.4]) by mx1.FreeBSD.org (Postfix) with ESMTP id A781843FCB for ; Thu, 10 Jul 2003 12:43:15 -0700 (PDT) (envelope-from max.clark@media.net) Received: from MCLARK (76.0.6.10.IN-ADDR.ARPA [10.6.0.76]) by Chow.corp.media.net (Netscape Messaging Server 4.15) with SMTP id HHTQL700.KE6; Thu, 10 Jul 2003 12:39:07 -0700 From: "Max Clark" To: "Dirk-Willem van Gulik" Date: Thu, 10 Jul 2003 12:47:39 -0700 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) In-Reply-To: <20030710180521.T96627-100000@foem> X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Importance: Normal cc: freebsd-hackers@freebsd.org cc: Dan Nelson Subject: RE: What ever happened with this? "eXperimental bandwidthdelayproduct code" X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 10 Jul 2003 19:43:16 -0000 First off, thank you everyone who has helped me with this. After the reading that I have been through this is basically what I've come to: throughput = window size / RTT Which means that if I want a 5Mbit/s throughput and my RTT is 220ms then I want a window size of 137,500 bytes. Is this inaccurate??? Besides these two sysctls are there any other changes that I should make to this system? net.inet.tcp.sendspace net.inet.tcp.recvspace Namely should I tweak/enable these sysctls? net.inet.tcp.msl (The Maximum Segment Lifetime for a packet.) net.inet.tcp.slowstart_flightsize (The number of packets allowed to be in-flight during the TCP slow-start phase on a non-local network.) net.inet.tcp.newreno (Enable TCP NewReno Fast Recovery algorithm, as described in RFC 2582) I started this thread with a discussion of these features, I am now under the impression that I do not want to enable/use these sysctls. net.inet.tcp.inflight_enable net.inet.tcp.inflight_debug net.inet.tcp.inflight_min net.inet.tcp.inflight_max net.inet.tcp.inflight_stab Thanks again everyone for their help! -Max -----Original Message----- From: Dirk-Willem van Gulik [mailto:dirkx@webweaving.org] Sent: Thursday, July 10, 2003 9:07 AM To: Max Clark Cc: Terry Lambert; freebsd-hackers@freebsd.org; Dan Nelson Subject: RE: What ever happened with this? "eXperimental bandwidthdelayproduct code" On Thu, 10 Jul 2003, Max Clark wrote: > > We've done this in the past for protocols such as ftp, http and smtp; and > it works wonders. > > The protocol is FTP, what sort of proxy are you talking about here? I would > like to have... Commercial/Custom written; mostly based on the apache framework. A B > ftp server freebsd proxy <---Network Link---> freebsd proxy > Windows PC > > I assumed that my first step in this configuration is going to be getting > the two freebsd boxes tuned and performing correctly. For that just follow: > http://www.faqs.org/rfcs/rfc2488.html And make sure you have met your requirements between A & B. Once that is done and confirmed working you can start thinking how to do the proxy things. There are quite a few ways to do it depening on what sort of use, user population, number of ftp sites behind the proxy, internet or not,etc. Dw.