On 04/03/2017 13:28, Michael Richardson wrote:
>
> Charlie Perkins <[email protected]> wrote:
> > I probably should have said more. I just meant that such nodes ought
> to be
> > allowed to have some network interfaces that do not participate in
> GRASP.
> > Please excuse the unclear brevity of my remark.
>
> Yes. generally, GRASP will be running over virtual interfaces provided by
> the ACP. The ACP is basically an overlay mesh using IP (maybe GRE) over
> IPsec transport mode using RPL to do routing. (IPsec Tunnel mode itself
> doesn't interact with routing protocols well. I wish I had an acronym for
> this very common usage...)
>
> So, nodes will have physical interfaces on which the ACP runs, and then
> ACP secured virtual interfaces on which GRASP will run. And then the limited
> subset GRASP will be used on the physical interfaces to discover new possible
> adjacencies.
Yes. Those details don't really belong in the protocol spec, but we will
adjust the wording because Charlie has pointed out that it's currently wrong.
Brian
_______________________________________________
Anima mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/anima
