>>>>> On Mon, 26 Jun 2000 09:57:26 +0200,
>>>>> Francis Dupont <[EMAIL PROTECTED]> said:
> I think it is not necessarily a better thing. A user might just
> want to specify a multicast group (with some zone ID) and let the
> kernel send packets to that group on an appropriate interface(s)
> according to the kernel's routing table, etc.
> => do you mean we should use the multicast routing table for
> genuine packets (BSD IPv4 multicast is not very clear about this
> for me, I have no concern but the specs need to be clear(er) about this).
Sort of that. We could install a route for ff00::/8 to some interface
as the multicast default route, or we could use a separate mechanism
to implement the default outgoing interface for multicast packets.
> This can make IPV6_MULTICAST_IF useless and/or things even more complex:
> what semantics for sin6_scope_id on the sending side do you propose
> (and is there a special case for the zero sin6_scope_id)?
I don't think such a default route makes IPV6_MULTICAST_IF useless,
but I admit it makes the situation complex. In any case, we should
clarify relationship and/or precedence about all these interfaces.
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
[EMAIL PROTECTED]
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
