At 08:32 20/07/2005, Brian E Carpenter wrote:
In reality /64 is an architectural boundary, even if in theory
it isn't. I don't believe that revisiting this is realistic. And I don't
believe it is in the least necessary.
Dear Brian,
i am afraid this is orthogonal:
- "In reality /64 is an architectural boudary", means it is because of the
practice, common thinking, etc.
- "I don't believe that revisiting this is realistic", means such
revisiting would come from a proposition.
You cannot impeach a new practice. One of the rules of network design is
"whatever is possible to try and is fun or brings money or worries to
someone will be used". But you can certainly propose ways towards a better
(network wise) support of such new practices. I suppose that you refer to
/64 in what we could name the "legacy" environment (the way the Internet is
today). Now think of a totally different approach of the Internet (we can
actually support 6 full ones with the /3 blocs) where a user could get far
less InterfaceIDs (and be happy with it) because for example the routing
system and ganularity would be different, or hardware would change, or NGN,
or MGN, or etc.
Up to now, what has been investigated is what IPv6 can bring to the
network. I think another interesting approach is to start from a universal
numbering space and investigate what IPv6 could bring to it (with the
current technology or not) and to all its constituents. May be the way to
understand how to deploy IPv6 faster instead of selling it slowly?
jfc
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
