Hi Peter, On Sat, 01 Nov 2008 17:03:44 -0700 "H. Peter Anvin" <[EMAIL PROTECTED]> wrote:
> Brian E Carpenter wrote: > > > >> I would not like to see IPv6 NAPT, but I see that as a real risk if > >> network is using DHCPv6 for allocating hosts just /128 addresses but at > >> the same time is not willing to delegate prefixes on demand. > > > > Agreed. > > The sad part is that it's probably inevitable given the decision to have > a fixed address length for IPv6. > I'm curious as to why you think this might be the case? >From what I understand, fixed length addresses were chosen for a few reasons (a) only CLNS had variable length addresses, verses every other protocol that didn't (e.g. applelalk, IPv4, IPX etc.), so there was much more limited experience with variable length addresses, (b) people tended to use fixed length CLNS addresses within their organisation anyway, to be operationally simpler and (c) it made building hardware forwarding ASICS etc. much simpler. I can't see why IPv6 having variable length addresses would have prevented people creating NAPT66 if /128s were allocated. Regards, Mark. -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
