Re: Host-based may be the way to go, but network controls are neccessary

[Kurt Erik Lindqvist <kurtis@kurtis.pp.se>]

> Future: when it becomes possible for a host to jump from one set of
> addresses to another (as in SCTP) it certainly becomes useful to do
> this. The crude way: drop packets, the host will get the hint. The nice
> way could be an ICMP message.
That means that the router would have to keep even more state than 
today so it would know when to send an update to the host.

State is among the most costly thing for network operators, you want to 
avoid it. To make this scale you would have to send periodic updates 
and the host would know only when the next update comes.

So with packet-loss or the above, there will first be a service impact 
- then a effect on the routing path. Now, if we assume that the primary 
reason for anyone to go to multihoming (I know there are others) is for 
network resilience (this will most likely be the first one at least), 
they are actually not getting that with the above model.

Why do I get that MPLS feeling here...

I realize that I am jumping in and out of discussions and not giving a 
very clear statement myself. I have promised both Sean and Michel to 
write up my impressions (FWIW) from Sundays meetings and post it to the 
list. I will try and get it done during the plenary..:)

- kurtis -