RE: modules of a mh solution (was RE: New multi6 draft: WIMP)

["marcelo bagnulo" <mbagnulo@ing.uc3m.es>]

> I have a suspicion that you do not mean "trial and error for discovering
> one's own list of addresses" but rather mean "trial and error as a basis
> for choosing one"...

Yes.

>
> I think there need to be mechanisms for both learning of one's own
> addresses and for helping decide which one to use.  The former is
> essential in the near-term, I believe.  Without it, we cannot scale any
> approach that makes the hosts aware of their multiple addresses.

I agree that the mechanism to learn one's own addresses is fundamental, but
i thought that the mechanism to do that would be stateless (or statefull)
address autoconfiguration...

Do you think that we may need something else?

>
> Something akin to routing table information, to help a host choose among
> the list of its own addresses, is important, but can be deferred.
>

I would say that you will need a mechanism to selects the destination
address the moment you want to be able to obtain fault tolerance by using
the available path. If not you may end up using a path that is not
available, i guess.


Regards, marcelo