CircleID

Yep.  It will be fun to see the rather in-grained hierarchial format of the IPv6 address meet the real world of non-hierarchial networks.  (Dual-homing, for instance.)

You are correct that we need "6to4 ''nat'' boxes", and frankly, we have for a while, but the pro-IPv6 crowd cannot allow themselves to think about those too much, even if they would speed IPv6 adoption.  The problem is if a useful "6to4" NAT was made, then one could do go through a "4to6" NAT from an IPv4 network, over an IPv6 backbone, then through an "6to4" NAT to get back to another IPv4 network.  On the face of it, such technology would speed up the adoption of IPv6 on the backbone, but conceptually it starts removing the need for "getting some more bits".  If you could solve the "6to4"/"4to6" NAT problems, you just stick two of them back-to-back, and you have solved the "4to4" NAT problems.

This seems to lead to the rabidly anti-NAT mentality, which is short-term depressing and long-term stupid.  The experience and concepts we are losing in forcing a single protocol, monolithic network are going to be costly in the long run.

...On the face of it, such technology would speed up the adoption of IPv6 on the backbone, but conceptually it starts removing the need for "getting some more bits".  If you could solve the "6to4"/"4to6" NAT problems, you just stick two of them back-to-back, and you have solved the "4to4" NAT problems.

Correct, and solving that problem in the absence of new functionality for IPv6 effectively means no need for IPv6.  (http://tools.ietf.org/html/rfc1669, from August 1994....)

/John

John Curran said:

Correct, and solving that problem in the absence of new functionality for IPv6 effectively means no need for IPv6.  (http://tools.ietf.org/html/rfc1669, from August 1994....)

Exactly.  Hence, this frequently leading to an anti-NAT mentality.