It seems to me that for multicasting a similar
kind of semantic overloading is currently common.
As for the unicast world, I don't think such
semantic overloading is helpful.
It seems to me that relative to multicasting,
there are at least 2 different objects that
matter:
1) A Multicast Group Identifier, which is topology
independent, and just gives a name for the group.
2) A Rendezvous Point useful for that Multicast Group.
I imagine this is topology-dependent, and aggregatable,
as it is naming the location of an RP (i.e. from the
perspective of the unicast world).
Some folks might suggest that (2) is a kind of "Forwarding
Directive" or "Forwarding Hint", but it seems to me that
the thing really needed is the location of a RP for a
given destination multicast group.
As well as a source address for a source-based distribution trees.
Check out one set of definitions in draft-farinacci-lisp-
multicast-01.txt.
Dino
Yours,
Ran
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
