Re: [armd] address resolution requirement from hosts to overlay edge nodes. Any opinion?

[Anoop Ghanwani <[e-mail]>]

On Mon, Feb 13, 2012 at 4:01 PM, Mike McBride <[e-mail]> wrote:

> On Fri, Feb 10, 2012 at 6:12 PM, Anoop Ghanwani <[e-mail]> wrote:
>> I think this is a very interesting problem.  If BOTH the MAC-to-IP binding,
>> and the MAC-to-tunnel-end-point binding can be resolved using a directory
>> service, it would eliminate the need for IP multicast support from the network,
>> provided there are no multicast applications.
>
> A directory could eliminate any reliance upon IP Multicast for MAC to
> tunnel end point binding but would not eliminate the need for IP
> Multicast in the network. You would still want the dynamic nature of
> joining and pruning along with the efficient data delivery across the
> overlay network. Good solutions will take mcast applications into
> proper consideration.

I agree that we would still need IP multicast in the network to
support multicast applications.  My point was that a directory-based
mechanism would eliminate the need for IP multicast if the only reason
it was need was for the purpose of getting ARP traffic through the network.

That said, there are unicast applications (such as Microsoft's Network
Load Balancing) that require multicast at L2, and because they
require multicast at L2, such a network would still need to support
IP multicast.

So really, the directory-based approach would be helpful
only in a narrow set of circumstances -- no multicast applications
including things like NLB.

Anoop