From owner-freebsd-ipfw@FreeBSD.ORG Wed Mar 3 15:04:49 2004 Return-Path: Delivered-To: freebsd-ipfw@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 9435C16A4CE for ; Wed, 3 Mar 2004 15:04:49 -0800 (PST) Received: from smtpout.mac.com (smtpout.mac.com [17.250.248.97]) by mx1.FreeBSD.org (Postfix) with ESMTP id 8692943D1D for ; Wed, 3 Mar 2004 15:04:49 -0800 (PST) (envelope-from cswiger@mac.com) Received: from mac.com (smtpin08-en2 [10.13.10.153]) by smtpout.mac.com (Xserve/MantshX 2.0) with ESMTP id i23N4nmC019113 for ; Wed, 3 Mar 2004 15:04:49 -0800 (PST) Received: from [10.1.1.193] ([199.103.21.225]) (authenticated bits=0) by mac.com (Xserve/smtpin08/MantshX 3.0) with ESMTP id i23N4mlf026336 for ; Wed, 3 Mar 2004 15:04:48 -0800 (PST) Mime-Version: 1.0 (Apple Message framework v612) In-Reply-To: <20040303120211.O8264-100000@oahu.WURLDLINK.NET> References: <20040303120211.O8264-100000@oahu.WURLDLINK.NET> Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <34868EFE-6D67-11D8-85AD-003065ABFD92@mac.com> Content-Transfer-Encoding: 7bit Cc: freebsd-ipfw@freebsd.org From: Charles Swiger Date: Wed, 3 Mar 2004 18:05:03 -0500 X-Mailer: Apple Mail (2.612) Subject: Re: ipfw/dummynet pipe size, is there a burst setting? X-BeenThere: freebsd-ipfw@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: IPFW Technical Discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 03 Mar 2004 23:04:49 -0000 On Mar 3, 2004, at 5:06 PM, Vincent Poy wrote: > ipfw pipe 1 config bw size > > to set the size of the pipe. I noticed on Linux, they can set a burst > size so it can burst a x number over the pipe size, is there a similar > setting available? I'm not sure I understand the semantics of this? How much bandwidth is available when using this burst mechanism-- and what traffic goes through when several things try to burst at once, for example? Anyway, you can use different weighted queues feeding into larger-sized pipe that will probably "do what you want"... > Also, I noticed for pipes, one can set the queue size > in slots of KBytes, how does one determine what's a good size? Consider the cross-product of network bandwidth times latency for your situation; a discussion of how to tweak the TCP receive window size is closely related... -- -Chuck