From owner-freebsd-questions@FreeBSD.ORG Sat Dec 24 15:58:50 2005 Return-Path: X-Original-To: freebsd-questions@freebsd.org Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 66A0616A41F for ; Sat, 24 Dec 2005 15:58:50 +0000 (GMT) (envelope-from danial_thom@yahoo.com) Received: from web33309.mail.mud.yahoo.com (web33309.mail.mud.yahoo.com [68.142.206.124]) by mx1.FreeBSD.org (Postfix) with SMTP id D49EF43D60 for ; Sat, 24 Dec 2005 15:58:48 +0000 (GMT) (envelope-from danial_thom@yahoo.com) Received: (qmail 72651 invoked by uid 60001); 24 Dec 2005 15:58:48 -0000 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:Received:Date:From:Reply-To:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=Z4+CxP/EkSj++JdjBMKQM1CVQaHiDbKArXzmneLiGY9w+mVhuVMoMGA+ZsyDXnS9jEoIXx3CStoxAGMlFa7782jH7jyFPSgpDUUNqsW4KP0BKRWzi7ESvNh8zKY/lcA8n8GC6I+A37b0ulg7hV/D6RDJ+2ETfdNum0xPMcqF0DM= ; Message-ID: <20051224155848.72649.qmail@web33309.mail.mud.yahoo.com> Received: from [24.46.186.215] by web33309.mail.mud.yahoo.com via HTTP; Sat, 24 Dec 2005 07:58:48 PST Date: Sat, 24 Dec 2005 07:58:48 -0800 (PST) From: Danial Thom To: Ted Mittelstaedt , "Winelfred G. Pasamba" In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Cc: pretenda@wrgn.net, freebsd-questions@freebsd.org Subject: RE: FreeBSD router two DSL connections X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: danial_thom@yahoo.com List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 24 Dec 2005 15:58:50 -0000 --- Ted Mittelstaedt wrote: > > http://www.edimax.com/html/english/products/PRI582.htm > > "...Performs Outbound load balancing by > session, weight round robin or > traffic..." > > Note that they say by SESSION not by PACKET. > > It's marketingspeak. They are simply using the > term load balancing > for a device that doesen't actually load > balance. Apparently > they figure that if they say "session load > balancing" even though > there is no such accepted definition, that then > they are somehow not > lying. > > It's akin to someone saying that "FreeBSD is a > kind of Linux" in a > sentence that uses Linux to indicate "open > source operating systems" > > Apparently you never heard the old saying "A > grain of truth is > buried in all great lies" I'm not sure what your primary language is, but "round robin" IS packet balancing. Suppose you have 2 "pipes": Round Robin: 1 packet to pipe1 1 packet to pipe2 1 packet to pipe1 1 packet to pipe2 Weighted round Robin, weighted 2 to 1: 1 packet to pipe1 1 packet to pipe1 1 packet to pipe2 1 packet to pipe1 1 packet to pipe1 1 packet to pipe2 "Per session" balancing may be useful when you have paths that are not very "equal". If you load balance to different ISPs packets could arrive out of order (in fact they are likely to). This is not really a problem for modern TCP stacks. Session balancing, if done properly, should guarantee that the ACKs for a download go out the same pipe as the data is arriving. Its not clear from the datasheet if thats the case, but thats the correct way to do it. Its seems like a quite comprehensive product to me, from the docs. Ted's analysis is backwards. "load balancing" is a vague term. "Weighted Round Robin" is a more specific term for how they have implemented the load balancing. Danial __________________________________________ Yahoo! DSL – Something to write home about. Just $16.99/mo. or less. dsl.yahoo.com