On Thu, 14 Aug 2003 13:37:56 +0100 Tim Chown <[EMAIL PROTECTED]> wrote:
> How different is this to the multi-addressing multihoming scenario, which > Christian Huitema's proposals are attacking? I think this is good work, > but the brickbats come from those who wish to push/enforce central policy > to hosts for the host address selection algorithms (moreso in larger > corporate networks than the SOHO networks that Christian's work seems very > well suited for). address selection won't work except in very simple cases. it's a lousy way of implementing routing policy. it needs to die, violently if necessary. but being able to renumber without breaking open connections (well, without breaking connections that are open with long periods of time) is really useful for disruption-sensitive environments. also my impression is that stateless address autoconf is a lot more generally applicable than DHCP, because it provides a stable address for the host without requiring any per-host configuration on the DHCP server. this makes renumbering easier because there's less configuration data that has to be changed. -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
