From owner-freebsd-hackers Wed Dec 6 13:28:01 1995 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id NAA22392 for hackers-outgoing; Wed, 6 Dec 1995 13:28:01 -0800 (PST) Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id NAA22381 for ; Wed, 6 Dec 1995 13:27:56 -0800 (PST) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id OAA01948; Wed, 6 Dec 1995 14:25:45 -0700 From: Terry Lambert Message-Id: <199512062125.OAA01948@phaeton.artisoft.com> Subject: Re: How big are queues on a typical router ? To: luigi@labinfo.iet.unipi.it (Luigi Rizzo) Date: Wed, 6 Dec 1995 14:25:45 -0700 (MST) Cc: terry@lambert.org, hackers@freebsd.org In-Reply-To: <199512061932.UAA18137@labinfo.iet.unipi.it> from "Luigi Rizzo" at Dec 6, 95 08:32:28 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-hackers@freebsd.org Precedence: bulk > > > I am asking because during some tests (on a connection which > > > apparently goes through a ~256Kbit/s line) I experienced large > > > packet losses, yet an upper bound on the RTT of about 1.2s compared > > > to a lower bound of ~.3s . This makes me think that the queue on > > > the routers is about 1s-worth of data (approx 32KB). Is this correct ? > > > > The RTT value you have is the minimum pool retention time for a > > successful transit of the router. > > > > This time dictates the number of packets, maximum, you can see in a > > given interval, and thus how much memory (what pool size) would be > > required. > > sorry I don't get this. If it takes 1.2s for a full transit, the amount of memory that a pool would need to contain to guarantee that the pool does not overflow is the number of packets that can be in the router at one time times the maximum packet size. > > It sounds like your router is simply overleaded by 20%. > > nor i can understand where the 20% comes from, how it relates to the > loss rate measured by a (20-minutes long) sequence of pings, and if > it is reasonable that this is a steady-state situation. You said an upper bound of 1.2s on the RTT and if the queue is 1s as you think, then 1.2/1.0 = 1.2 = 120% capacity = overloaded by 20%. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.