> From: Keith Moore <moore(_at_)cs(_dot_)utk(_dot_)edu>
> NATs do harm in several different ways
It's not just NAT's that are a problem on the fronts you mention, though:
> they block traffic in arbitrary directions
My ISP blocks incoming SMTP and HTTP connections. Has nothing to do with
> these days they often destroy transparency.
Some ISP's trap outgoing HTTP requests and silently divert them to caches.
Again, it's not just NAT that's doing this.
> NATs started with a simple design, pretended it would work well
> without doing the analysis,
Actually, I think the people who started NAT's (mostly Paul T) understood
quite well what the problem were going to be. It's just that NAT was such
a simpler/cheaper solution in the short term that it was too attractive.
Realistically, the last chance to avoid NAT was when variable-length
addresses were removed from IP somewhere in the TCP 2.5 -> TCP 3.0 -> TCP
3.1 transition (I don't know exactly which stage it was). In other words, a
*loooonnnnggg* time ago. We've just been along for the ride ever since.
Ietf mailing list