In einer eMail vom 23.03.2008 16:18:35 Westeuropäische Normalzeit schreibt
bill@herrin.us:
The new protocol will have to be dual-stack for a You are so right, Bill, by pointing out, that the solution must
be dual-stack.(and if there were even several solutions, then multi-stack). I
like to go further and say, every solution must be incrementally deployable
as it has to deal with a growing/changing internet even after being well
deployed (like BGP which today still must be and is incrementally
deployable).
I can't imagine a concept which needs a flag day, and if any one came up
with such a concept he easily could be helped.
I am very much in favor of replacing BGP by some better solution, but for
the start BGP is a big help:
The incremental deployment of a new protocol does not have to start from
one single point in the network nor is it necessary that it expands along strict
links from there.
BGP-based VPN (RFC2547 I think) provide some fundamental
discovery mechanism for combining remote parts as to form some new
structure (there the VPN tunnel mesh). It can be re-used as a starter
motor.
Hence, I do not like the current discussion by which the incremental
deployability argument is used for better qualifying the own resp. disqualifying
any other solution.
Heiner
|