Re: [homenet] draft-boutier-homenet-source-specific-routing-00

[Juliusz Chroboczek <[e-mail]>]

> Y is just announcing that it only does forward normal traffic, its not
> part of the "source-specific" topology.

How do you identify "normal traffic"?

Recall that we're not "tagging" source-specific packets -- these are
perfectly ordinary IP packets that happen to have a source address
that matches a source-specific route.

(If you start tagging packets with the gateway's address, then you're
doing weak source routing -- which is a rather cool technology in
itself, see reference [CLARK] in draft-boutier-tralala, but it is out
of scope for this particular work.)

> I think its the same issue/misunderstanding we had in the last post.
> The question if traffic for an unspecific source-address can go out of
> any gateway or not. If it can go out of any gateway, A will just take
> the traffic and forward it to the external gateway without asking Y.

Yeah, that's essentially equivalent to redistributing a non-specific
default route from each gateway, perhaps a blackhole route, and
perhaps with a high metric.  I guess that's what we need to do, since
it breaks routing loops in many cases.  It makes me vaguely uncomfortable,
but I'm not quite sure why.

-- Juliusz