Re: IPv6<->IPv4 address translation is not NAT?

[Geoff Huston <gih@apnic.net>]

Ralph Droms wrote:
> Geoff - in the v6ops meeting you said that the IPv6 transition problem 
> is not a NAT problem.  Can you explain in more detail?  There must be, I 
> think. a translation between addresses from incompatible address spaces, 
> just like NATv4.  What do you think is the difference?


The point I was trying to make is that NATs are not "just" Network 
Address Translators - they are in fact used for a particular style of 
Network address translation which is asymmetric. We all know how a NAT 
is defined - it has an "inside" and an "outside" and packets from the 
"inside" that are heading to the "outside have one set of transforms 
applied to the IP source field and packets in tyhe opposite direction 
have a comparable set of transforms applied to the IP destination field.

So we have assumptions of "inside" and "outside", and only one part of 
the header is touched, etc etc.

The point I was making at the microphone in the WG was that the V4 / V6 
translation problem shares very little with the asymmetric NATS that we 
know as "NAT" as this protocol transform is an entirely different form 
of address translation which has different triggers, semantics and 
states. And I was trying to say that calling them both "NATS" is 
uncomfortable for me.

Now is that clearer or muddier?

   Geoff