[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Draft of updated WG charter
On Fri, Jan 09, 2004 at 01:37:24PM -0800, Tony Li wrote:
> Well, then I would pose to you the obvious question: what is
> it that you want multihoming to do? If the end host behavior
> does not change, then all you are left with changing is the
> routing subsystem. And the only thing that you can now do
> is to declare all multihomed prefixes global or do an absurd
> amount of tunneling to repair the topology. For obvious reasons,
> I hope we agree that both of those alternatives are even less
Tony, if it was an easy answer, we wouldn't be having this discussion.
Whatever solution we come up with must somehow account for the realities
on the ground. I can see there being multiple methods in use, with
people hiding deep inside networks willing to accept the fact
that their sessions may not survive rehoming.
IFF we can get a solution working as transparently as say, tcp/ip
congestion control or dns, then we have a shot I'd think.