Date: Thu, 10 Jul 2003 09:06:34 -0700 From: "Max Clark" <max.clark@media.net> To: "Dirk-Willem van Gulik" <dirkx@webweaving.org>, "Terry Lambert" <tlambert2@mindspring.com> Cc: Dan Nelson <dnelson@allantgroup.com> Subject: RE: What ever happened with this? "eXperimental bandwidthdelayproduct code" Message-ID: <ILENIMHFIPIBHJLCDEHKCEHFCJAA.max.clark@media.net> In-Reply-To: <20030710115853.O96627-100000@foem>
next in thread | previous in thread | raw e-mail | index | archive | help
> 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... ftp server <LAN> freebsd proxy <---Network Link---> freebsd proxy <LAN> Windows PC I assumed that my first step in this configuration is going to be getting the two freebsd boxes tuned and performing correctly. Thanks, Max -----Original Message----- From: Dirk-Willem van Gulik [mailto:dirkx@webweaving.org] Sent: Thursday, July 10, 2003 3:04 AM To: Terry Lambert Cc: Max Clark; freebsd-hackers@freebsd.org; Dan Nelson Subject: Re: What ever happened with this? "eXperimental bandwidthdelayproduct code" On Thu, 10 Jul 2003, Terry Lambert wrote: > > Okay, let's say how do I force my machine to think it doesn't have any > > latency and saturate a 6Mbit/s link even though the link has 220ms latency? > > See the recent discussion on the FreeBSD-performance mailing list. Your propblem is similar to that encountered in sat links; where you have at least a 2 x 280ms RTT's and exteremely reliable/error free 'big' links (ok, that is not quite true; but the ECC is configurable so you simply set the error rate; and use power, width and dish size amongst other as a design parameter to play iwth). Another goed overview document is best current practices for satellite links (BCP28): http://www.faqs.org/rfcs/rfc2488.html or http://mmlab.snu.ac.kr/course/AdvancedInternet/reading/SatelliteTCP.htm Alternatively if a specific protocol is involved the use of a proxy for that protocol to intentionally break end to end semantics can do wonders. We've done this in the past for protocols such as ftp, http and smtp; and it works wonders. Dw
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?ILENIMHFIPIBHJLCDEHKCEHFCJAA.max.clark>