Marcelo, Iljitsch, what was the method you used to "disable v4"?
I agree with Dave that an important scenario to consider for this particular functionality is a host without any v4 stack; but since there aren't many of those its going to be hard to test this. In some sense we can rely on future implementors to do the right thing, particularly if we publish documents saying what the right thing is.
However, another important scenario that I would like to test is having a normal host (say, Vista or Linux) that can support both versions, and then running that in a network that has no IPv4. That is, DHCPv4 does not respond, routers do not route IPv4 packets, etc. (The host might come up with some link local addresses, however.) In that environment, what happens? If typical implementations fail in some manner in this scenario, it means that the NAT-PT replacement is going to have a hard time in any network except those involving purpose built new hosts...
Jari _______________________________________________ Int-area mailing list [email protected] https://www.ietf.org/mailman/listinfo/int-area
