From owner-freebsd-questions@FreeBSD.ORG Tue Sep 13 19:24:15 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 1F1E616A41F for ; Tue, 13 Sep 2005 19:24:15 +0000 (GMT) (envelope-from cswiger@mac.com) Received: from smtpout.mac.com (smtpout.mac.com [17.250.248.45]) by mx1.FreeBSD.org (Postfix) with ESMTP id BAD8443D45 for ; Tue, 13 Sep 2005 19:24:14 +0000 (GMT) (envelope-from cswiger@mac.com) Received: from mac.com (smtpin02-en2 [10.13.10.147]) by smtpout.mac.com (Xserve/8.12.11/smtpout11/MantshX 4.0) with ESMTP id j8DJOEpV024093; Tue, 13 Sep 2005 12:24:14 -0700 (PDT) Received: from [10.1.1.209] (nfw2.codefab.com [199.103.21.225] (may be forged)) (authenticated bits=0) by mac.com (Xserve/smtpin02/MantshX 4.0) with ESMTP id j8DJOCXZ028658; Tue, 13 Sep 2005 12:24:13 -0700 (PDT) In-Reply-To: <20050913152320.27919.qmail@web33305.mail.mud.yahoo.com> References: <20050913152320.27919.qmail@web33305.mail.mud.yahoo.com> Mime-Version: 1.0 (Apple Message framework v734) Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed Message-Id: Content-Transfer-Encoding: 7bit From: Charles Swiger Date: Tue, 13 Sep 2005 15:23:59 -0400 To: danial_thom@yahoo.com X-Mailer: Apple Mail (2.734) Cc: freebsd-questions@freebsd.org Subject: Re: VLAN interfaces on FreeBSD; performance issues X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 13 Sep 2005 19:24:15 -0000 On Sep 13, 2005, at 11:23 AM, Danial Thom wrote: > its not clear why Chuck keeps answering since he > clearly doesn't understand the question. I'm willing to try and help people, even if the questions being asked aren't entirely clear. If you want to believe this reflects a lack of understanding on my part, that's OK: you're welcome to hold that opinion. > You can, of course, multihome with one nic, and > Spanning Tree and "collision domains" have > nothing to do with anything, simply by routing to > the correct router. A machine with one NIC can be attached to a "multihomed network". But a machine with one NIC is not a "multihomed machine". [1] > "Multi-homing" refers to having more > than one network egress (ie 2 or more upstream > providers) and the ability to "decide" which one > to send specific traffic to. Sure. This definition of "multihoming" is applicable to the network as a whole, not to each and every individual device on the network. A multihomed network can lose one of its upstream connections and still retain full connectivity, because there is an alternate path available via the second (or additional) upstream connections. In order to construct such a network, you need two or more routers, each of which is a "multihomed machine" by the classic definition (ie, has two physical network interfaces connected to two different physical networks), and you commonly use BGP to coordinate routing with the upstream providers, just as you might use VRRP or CARP to provide a single fault-tolerant virtual router IP for the systems on the LAN which will continue to function even if one of the routers fails. -- -Chuck [1]: I am aware that some people would disagree with this. For example, Microsoft's IIS documentation apparently describes a webserver hosting more than one domain as "multihomed" rather than using Apache's terminology of "name-based virtual hosts". There are people who believe that using "ifconfig alias" to configure additional IPs on a NIC creates a multihomed system, but there is no physical redundancy involved and there is no isolation of traffic. I find such usages of the term "multihomed" to be misleading at best, and at worst sometimes even represent a deliberate effort to confuse people expecting the additional reliability and redundancy of a truly multihomed network architecture: What happens to a machine with a single NIC when that NIC fails? Do you see any difference between this and a machine with two or more NICs? (The latter retains network connectivity, the former does not.)