Re: Shim6 proxies

[Scott Leibrand <sleibrand@internap.com>]

On 04/20/06 at 5:08pm +0300, marcelo bagnulo braun <marcelo@it.uc3m.es> wrote:

> El 20/04/2006, a las 16:58, Scott Leibrand escribió:
>
> > As I've said before, I think the shim6 design needs to recognize that
> > it will not be the One and Only method for multihoming, and therefore
> > it needs to be designed to ensure that hosts that don't use shim6 for
> > multihoming can still interoperate with multihomed hosts and small
> > sites that do want to use shim6.
>
> I am not sure why do you think this is not the case with this
> particular aspect...

To some degree it is, I'm just worried that the "everyone will want to
shim6" assumption is clouding our vision when it comes to minimizing the
burden on hosts that don't use shim6 for multihoming.

> I mean a big site that is multihoming using bgp will not need to do
> source address based routing of course, not even in the case that it is
> communicating with a smallish site that is using the shim and that it
> does source address based routing...

Right.  I'm referring to the discussion around "how much state is required
on busy servers", and "how much of that state can be safely discarded".
An important design goal IMO needs to be to make sure we don't require any
more resources than absolutely necessary at the non-multihomed end to make
shim6 work.

Perhaps another way to look at it is considering requirements for
"shim6-aware" hosts, which wouldn't multihome with shim6, but would talk
shim6 to multihomed correspondents.

An alternative, of course, is going back to the original topic of this
thread, which is designing the protocol to allow offloading of the shim6
work and state-tracking to an external proxy.  This is also something I
think will be important in getting enough content providers "shim6-aware"
to make shim6 deployment useful for multihomed hosts and small sites.

-Scott