Hi Dan,
So your position is that I must do one of those two rather than using site-local addresses on the one segment along with the two globals? I believe that this restriction renders site-locals useless for many applications. I would be forced to opt for NAT.
As far as I can tell, Michel's view on this does not represent the majority view of people who favour the use of site-local addresses for private addressing.
The scoped addressing architecture currently assumes that there will (potentially) be global and site-local prefixes in use on the same subnet. Due to the way that we expect IPv6 address to be allocated (a /64 per subnet) and assigned (using address autoconfiguration), though, there will be a tendency for all of the nodes on a single network to use the same set of prefixes. If a global prefix is advertised on the subnet, all nodes on that subnet would configure global addresses. Obviously, this could be overridden using manual configuration or by using DHCPv6 for address assignment. Margaret -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
