From owner-freebsd-hackers Wed Jan 17 16:14:52 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id QAA20788 for hackers-outgoing; Wed, 17 Jan 1996 16:14:52 -0800 (PST) Received: from underdog.maxie.com (maxie.com [199.250.231.28]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id QAA20783 for ; Wed, 17 Jan 1996 16:14:43 -0800 (PST) Received: (from max@localhost) by underdog.maxie.com (8.6.12/8.6.12) id TAA18370; Wed, 17 Jan 1996 19:13:07 -0500 Date: Wed, 17 Jan 1996 19:13:06 -0500 (EST) From: James Robertson To: Nik Malenovic cc: freebsd-hackers@freebsd.org Subject: Re: Another cool hack with FreeBSD... In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-hackers@freebsd.org Precedence: bulk On Wed, 17 Jan 1996, Nik Malenovic wrote: > ISDN cards can use both channels but it's done via hardware. bridges > route ethernet traffic based on MAC addresses (hardware solution). Okay, I was assuming the ISDN cards had some routing capability of thier own, like the external ISDN unit here. I've personally never met one of the internal beasts. > It would be interesting to see load balancing being standardized. > you can load balance ANY interface. let's say a device in kernel > to which you add multiple interfaces, which are multiple lines, > with the same routing info and kernel knows it can send packet > via any of them.. Any ideas how Linux and CISCO do 'em? I really do not know much about how FreeBSD handles IP at the kernel level, but it would seem to me something like that could be handled at the route layer... Instead of just disgarding packets it could check for an alternate route to the same destination. With a little extra accounting (keeping track of link speed and current percentage of use) it could be rather efficent at delivering packets by the best link at that particular moment. James Robertson Treetop Internet Services