On Sat, 2002-11-23 at 23:43, Michel Py wrote:
> > Mika Liljeberg wrote:
> > This means that what we really need a site-local
> > prefix autoconfiguration specification similar to
> > the link-local address autoconfiguration.
>
> This makes little sense. Site-locals involve routing between subnets and
> router configuration, this falls into the categories of things that
> still requires a network administrator to do. The point that site-locals
> addresses are *not* automatically configured has been made several times
> here.
Yes, I know that there are cases where the network admin will want to
configure the subnets manually. I'm not suggesting that site local
prefixes would be autoconfigured by default. This would have to be
enabled per router-interface by the admin.
Still, I think that there is something to be said for being able to put
together a routed site-local network with the same level of convenience
as sticking together a bunch of switches.
Why not extend Savola's idea to also generate pseudo unique /64's for
leaf subnets? A simple duplicate prefix detection mechanism could be
added to make it reliable (or the IGP could handle it). This would still
provide address stability and retain the possibility to combine or
inter-connect separate site-local networks.
MikaL
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
