Re: NATmakes a network a host -- must every process have an IP address?

This is a sloppy reply to something I saw below.  I apologize to 
threaded mail readers that will misplace this message...;)

>Graham> Where this all leads, I think, is that the worst thing about NAT is
>Graham> that it hinders the deployment of new applications.

I was on a call today where we went over the problems of adding 
DNSSEC to DNS in the face of what amounts to NATs.  The following 
point was made:

The problem that is appearing is not due to DNSSEC, nor due to DNS, 
nor really due to NAT.  The problem is rooted in that DNS was not 
designed to run across NATs - and the implication is that forward 
progress (extensions) is hindered because of this.

I'm not convinced that NATs 'hinder' the deployment of new 
applications, in the sense that 'hinder' means 'prevents,' or 'stops 
cold.'  The presence of NAT does call for a more sophisticated 
protocol (okay, complicated), I'll grant you that.  (I should add 
that I may be naive here.)

I'm convinced that had DNS built in features to remotely manage 
answers from a cache (which is the crux of the problem - a cache 
sitting topologically on a NAT node) and had made error reporting 
more explicit, DNSSEC wouldn't be facing the problem de jure.

So, I'm convinced that NAT hinders extension of existing (pre-NAT) 
applications.
-- 
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Edward Lewis                                          +1-703-227-9854
ARIN Research Engineer

Received on Thursday, 5 December 2002 18:38:12 UTC