On Thu, Jul 12, 2012 at 3:40 PM,  <[email protected]> wrote:

> Actually I didn't intend to go there (at least not in this thread), as
> DHCP traffic is L2 LAN/VLAN scoped, whereas VDP (and ECP) are scoped to
> a single L2 link, and hence have rather different requirements/behavior
> in terms of where they hit the control plane.

That scope is the case as defined today, but that can
be changed.  We can always define the capability for
ECP/VDP to be sent to a specific unicast address.
There is a precedent for that with LLDP & LACP.
However, we then have to address the problem for how
to discover that unicast address.

> The only thing I'm trying to do is reuse the words "relay" and "proxy" as
> they are used with DHCP to differentiate two classes of possible VDP
> functionality in a bridge between an End Device and an NVE.  A "relay" just
> forwards the traffic associated with the protocol whereas a "proxy" is a
> full participant in the protocol.

This type of operation is less clear to me.

Anoop
_______________________________________________
nvo3 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nvo3

Reply via email to