From owner-freebsd-questions Tue Mar 14 22:04:20 1995 Return-Path: questions-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id WAA11300 for questions-outgoing; Tue, 14 Mar 1995 22:04:20 -0800 Received: from zap.zap.qc.ca (zap.zap.qc.ca [198.168.127.8]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id WAA11294 for ; Tue, 14 Mar 1995 22:04:17 -0800 Received: (from fortin@localhost) by zap.zap.qc.ca (8.6.9/8.6.6) id BAA02907; Wed, 15 Mar 1995 01:03:47 -0500 From: Denis Fortin Message-Id: <199503150603.BAA02907@zap.zap.qc.ca> Subject: Re: Routing and Ethernet To: ugen@netvision.net.il (Ugen J.S.Antsilevich) Date: Wed, 15 Mar 1995 01:03:47 -0500 (EST) Cc: jg@euronet.nl, questions@FreeBSD.org In-Reply-To: from "Ugen J.S.Antsilevich" at Mar 14, 95 05:46:54 am X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Content-Length: 1374 Sender: questions-owner@FreeBSD.org Precedence: bulk > >You cannot have two physically different wires that have the same network > >(193.78.175.*) on them. This is a TCP/IP issue, not a FreeBSD problem. > > Wrong! You can and how...as a simple example i have on my machine two adresses > ep0 is 194.90.1.15 and ed0 is 194.90.1.18... Hmmm. I could see that this might be possible if both cards are on the same physical Ethernet (though I can't really see how that would be useful unless you've got a slow ethernet card). I cannot see how it would make sense if both cards are on different Ethernet segments and have the same netmask (0xffffff00)!!! In that case, there would be no way of determining through which interface (and hence on which wire) a packet destined to 194.90.1.17 would go. > If you can have two different IP's for same interface you even better can > have different IP's for different interfaces no matter on which network... I'm not sure I follow this. If I understand right, it seems to violate half a dozen RFCs and create difficult routing problems. With most TCP/IP implementations I've seen, it is possible to build fully working but incorrect network configurations. Sometimes it may even (appear to) be useful to do this. -- Denis Fortin fortin@acm.org DMR Group Inc, (514) 877-3301 These opinions are my own