Date: Fri, 19 Jul 2002 23:45:44 -0400 (EDT) From: Garrett Wollman <wollman@lcs.mit.edu> To: Rik van Riel <riel@conectiva.com.br> Cc: freebsd-net@FreeBSD.ORG Subject: Re: Another go at bandwidth delay product pipeline limiting for TCP Message-ID: <200207200345.g6K3jise017051@khavrinen.lcs.mit.edu> In-Reply-To: <Pine.LNX.4.44L.0207200005190.12241-100000@imladris.surriel.com> References: <200207200245.g6K2jHOh081549@apollo.backplane.com> <Pine.LNX.4.44L.0207200005190.12241-100000@imladris.surriel.com>
next in thread | previous in thread | raw e-mail | index | archive | help
<<On Sat, 20 Jul 2002 00:12:46 -0300 (BRT), Rik van Riel <riel@conectiva.com.br> said: > Absolutely. On my DSL things work best when I limit myself > to 220 kbit/s down (from the maximum 256 kbit/s), or about > 85% of the bandwidth. That is hardly surprising. In fact, that is a well-known result of Queueing Theory. It is impossible to achive better than 85-90% utilization in the presence of appreciable non-scheduled traffic competing for the same channel. (That's a big part of why protocols like ATM were designed to require everything to go through admission control and shaping: otherwise you can't make tight guarantees of latency and drop rates while still maximizing channel utilization.) -GAWollman 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?200207200345.g6K3jise017051>