On 2 dec 2008, at 20:47, Hallam-Baker, Phillip wrote:
If you look at the strategy for the end-of-life for IPv4, surely it
will involve placing every network that is not already behind a NAT
behind NAT64? As in, the day that IPv4 addresses are no longer being
maintained by the NICs and the backbone routers are refusing to
accept BGP traffic advertising v4.
No, that wouldn't work. If hosts still have v4, then routers still
need to have those prefixes in their tables to get the packets
delivered. So if you remove the v4 prefixes from routers, you need to
start routing the IPv6 versions of those prefixes, importing the IPv4
routing table swamp into the IPv6 routing table, where it takes up
more room because the addresses are longer. Not a good idea.
What will happen is that as IPv4 traffic becomes a smaller part of the
total traffic, IPv4 routing will be relegated to older boxes. Just
like IPv6 routing started on cheaper, separate routers.
You don't really put v4 boxes "behind" NAT64, as NAT64 allows sessions
in the v6->v4 direction. Rather, you put the IPv4 internet "in front"
of NAT64.
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf