On 28-Feb-19 15:01, Michael Richardson wrote:
>
> Brian E Carpenter <[email protected]> wrote:
> >> Brian E Carpenter <[email protected]> wrote:
> >> > Here is a new draft for ANIMA to consider. We believe it is useful
> >> > to explore the idea of an autonomic control plane constructed at
> >> > the link layer. Comments please!
> >>
> >> Does this mean it transports L2 packets (a la spanning-tree or TRILL),
> or
> >> does this mean it uses L2 technologies like MACSEC to create a tunnel
> >> for L3 packets?
>
> > Well, I don't think the draft in its present form should attempt to
> > answer that question. In my mind I was thinking of a conventional
> > LAN set up, i.e. spanning tree, but that doesn't have to be the
> > only way. In practice that usually needs MLDv2 snooping as well.
>
> okay, so you were thinking about the first part.
> Yes, you need MLDv2 snooping.
>
> >> i.e. does L3 multicast appear to just work because it more layer-2
> tricks?
>
> > L3 multicast works automatically if L2 multicast works, surely
> (remembering
> > that there is no router). How L2 multicast works is another question, of
> > course. I don't mean to say it's unimportant, just that it's not in
> scope
> > for this particular draft.
>
> In other words, more IPv4-inspired L2-tricks to maintain the illusion there
> is a big-blue cable with AUI taps on it. And continued inability to see L2
> switches, or creatively route around L2 failures :-)
I'd be happy enough to see that done too, but that's a much bigger step
than what we're suggesting, which is a way to get the ANI deployed
in smaller shops than will run a full BRSKI/ACP setup.
As for emulating classic Ethernet, yes, that is how the world works
today in many places.
Brian
_______________________________________________
Anima mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/anima
