From owner-freebsd-net@FreeBSD.ORG Wed Sep 12 02:31:46 2007 Return-Path: Delivered-To: freebsd-net@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id ADBEF16A421 for ; Wed, 12 Sep 2007 02:31:46 +0000 (UTC) (envelope-from ivsan@ngs.ru) Received: from imx3.ngs.ru (imx3.ngs.ru [81.176.214.138]) by mx1.freebsd.org (Postfix) with ESMTP id 3CCEF13C461 for ; Wed, 12 Sep 2007 02:31:46 +0000 (UTC) (envelope-from ivsan@ngs.ru) Received: from mx1.intranet.ru (mx1.intranet.ru [172.16.1.1]) by smtp.ngs.ru (smtp) with ESMTP id C6BC42DB0148 for ; Wed, 12 Sep 2007 09:31:43 +0700 (NOVST) Received: from [80.242.64.3] (account ivsan@ngs.ru) by mx1.intranet.ru (CommuniGate Pro WebUser 4.3.12) with HTTP id 95372950 for freebsd-net@freebsd.org; Wed, 12 Sep 2007 09:31:43 +0700 From: "Ivan Alexandrovich" To: freebsd-net@freebsd.org X-Mailer: CommuniGate Pro WebUser Interface v.4.3.12 Date: Wed, 12 Sep 2007 09:31:43 +0700 Message-ID: In-Reply-To: <46E58478.3070009@elischer.org> References: <46E58478.3070009@elischer.org> MIME-Version: 1.0 Content-Type: text/plain; charset="KOI8-R"; format="flowed" Content-Transfer-Encoding: 8bit X-Anti-Virus: Kaspersky Anti-Virus for Sendmail with Milter API 5.6.20, bases: 20070912 #412467, check: 20070912 clean Subject: Re: nested vlans and ethernet frame size X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 12 Sep 2007 02:31:46 -0000 Thanks for all your replies and explanations. Sure lowering mtu below 1500 bytes for all hosts or using jumbo frames would be the right solution especially for corporate network. But we probably have to stick to 1500 bytes for one reason: that FreeBSD-based router configuration I mentioned is for segments where we use seperate vlan for each client connection (we're an ISP): it seems that using anything else instead of 1500 for clients will require too much tech support efforts. Since double-tagged frames will be transmitted only by one ethernet link per segment (the link between upper level switch with QinQ and FreeBSD router) then maybe it's acceptable to allow a little deviation from the standard as long as we remember the implications. Thanks for the help Ivan