Masataka Ohta wrote:
I have written an Internet Draft to explain end to end NAT. http://tools.ietf.org/html/draft-ohta-e2e-nat-00
You can see the only reason to deploy IPv6 to keep the freedom of end to end transparency is now non-exsitent.
So, to keep IPv4 until we are ready with something much better than IPv6, why not mandate some form of NAT, legacy, end to end or whatever.
Masataka Ohta
PS
First thing we should do is to make initial PA allocation /24 and reduce the number of IP addresses allocated to an end user by 1/256 or so.
Please say you don't really mean this and it's just a joke you're playing on the Internet community ... NAT has been a kludge from the beginning, and now you try to "fix" the low availability of IPv6-capable hardware (which is finally starting to pick up a bit) by implementing yet another, even worse kludge? And what's that, "ready with something much better than IPv6"? Heck, it took much too long for working v6-equipement already, with even large vendors not having implemented it completely and reliably ... you mean to tell folks now to flush 10+ years of work down the drain, just because you have the "miracle cure" against (temporary) IPv4 shortage, and we now have another 10-20 years until we _REALLY_ run out of IPs? (please check on the time it usually takes standards committees to pass something like a new protocol, e.g. how long it took to get IPv6 standardized ...) Also, I don't see where your E2E really fixes reachability issues that current NAT has. Sure, it may fix the multiple-port problems of current NAT (which is already fixed by decent firewalls) Apart from that, your draft requires changes in both the gateways AND the applications --- do you honestly believe that _that_ can be implemented decently before IPv4 exhaustion? But maybe I'm just too stupid to see the genius behind your proposal ... -garry