Re: survivability, rewriting

["Spencer Dawkins" <spencer@mcsr-labs.org>]

Just a minor comment:

----- Original Message ----- 
From: "Iljitsch van Beijnum" <iljitsch@muada.com>
To: <mbagnulo@ing.uc3m.es>
Cc: "Multi6 Mailing List" <multi6@ops.ietf.org>
Sent: Tuesday, November 04, 2003 6:29 AM
Subject: Re: survivability, rewriting
>
> That is a possibility, but on the other hand rehoming in the blind
also
> has its risks. For instance, if an upper layer protocol provides too
> many hints, we could be rehoming when there was just some incidental
> packet loss. I'm pretty sure rehoming will have to trigger slow
start
> in TCP, which is not something you want to happen if isn't
absolutely
> necessary.

In some unrelated conversations in TRIGTRAN, we were saying that Slow
Start is conceptually the right thing to do for path changes, but Slow
Start works fine if you lose something - why not wait to see what you
lose, and only Slow Start if you need to?

If you lose packets during rehoming, you'll fast retransmit/fast
recover at a minimum, so it's not like you're switching paths at the
same sending rate.

But this is probably a good TSVWG question.

Spencer