> the conceptual data structure says that there's default router list
> and neighbor cache, not routing table.
The destination cache is essentially a routing table. Rather than
getting built up from routing updates recieved via routing protocols,
it is built up on demand by:
1) when communication with a destination first starts. If no
destination cache entry for that destination exists, it gets
created. Later, should a redirect for that destination arrives, it
updates that entry.
2) the default router list is exactly that. A list of default routers
one can pick a router from to use in those cases where one doesn't
already have a working router in the destination cache.
> from RFC2461 5.1 (conceptual data structures) I feel that the authors
> assumed that:
> - autoconfigured hosts are at the very edge of the site only
No such assumption was made by this author.
> - if we put a host onto a link with more than 2 routers, they will not
> automatically be configured, and do something tricky (like running
> routing daemon on hosts in receive-only mode).
Not at all. They can configure themselves just as if they are on a
leaf network. Redirects correct the situation where the "wrong" router
is chosen for a particular destination.
Thomas
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
