Randy Bush wrote:
And the NAT-PT implementation at NANOG (naptd) did seem to work once some configuration issues were ironed out. Unfortunately, this was not resolved until the very end of the meeting.
your made heroic efforts with the linux nat-pt, and finally got it. but do you think it will scale well?
For the size of a NANOG meeting, it seemed to be sufficient. I don't think I'd recommend trying to put thousands of users behind it though.
i suspect that all the nat-pt implementations are old and not well maintained. this needs to be fixed.
Still trying to understand deployment scenarios for nat-pt. I could see a case for very controlled environments with uniform clients (with robust v6 support). Outside of that, native-v6 + v4-nat (as outlined in Michael Sinatra's lightning talk) and Alain Durand's v4v6v4 seem more likely deployment candidates. That said, nat-pt is very useful for exercising native v6 support in clients and their applications. -Larry