Re: survivability, rewriting

[Masataka Ohta <mohta@necom830.hpcl.titech.ac.jp>]

Spencer;

> In some unrelated conversations in TRIGTRAN, we were saying that Slow
> Start is conceptually the right thing to do for path changes,

Agreed.

However, path changes at the middle of the path are not noticable
by end systems.

Considering the diameter of the Internet and assuming that reliability
of backbone and local ISPs are same, such path changes are expected
to occur more often than rehoming.

In addition, rehoming is a subset of path chang.

So, we, multi6, are not required to address the issues related to
path changes.

Right?

> But this is probably a good TSVWG question.

How about let routers accumulate (with XOR or something like
that) link CRC (or some psuedo random number involving link
specific numbers such as MAC addresses of a hop) into (so far
useless) flow label field to let upper layers notice the path
change to trigger the transport layer?

I myself am afraid that it is too late to change IPv6 router
behaviour. So, I never proposed any multihoming solution involving
router changes.

However, I know transport people are actively proposing to change
router behaviour to help TCP and they have better chance with IPv6. :-)

						Masataka Ohta