RE: issue 11 -- window size

Geoffrey Huang writes:
> The only hang-up I have is Jari's question below; namely, what exactly
> constitutes "information needed to do mobility 
> and path testing"?  It has to be fairly lightweight since the expectation is
> that this information would be included in each packet.

Depending on the NAT issue, it could even be simply the IP-addresses
already there in the outer header. Especially if we do not do
automatic updates based on the path tests, but after we detect from
the path test that the IP-addresses pair has changed, we do actual
notification of movement as a separate exchange (which still gets the
same processing, so the packets will get trough and exchange ends even
when change happens again during that exchange).

Perhaps the largest thing that could be there would be list of
IP-addresses peer has, but I am not sure we need that. 
-- 
[email protected]