RE: [BEHAVE] Comments on the NAT66 draft

["Wes Beebee (wbeebee)" <wbeebee@cisco.com>]

I guess RFC 4864 doesn't go quite far enough - it says (paraphrasing): 
You shouldn't need NAT66 because there are other ways to accomplish your
goals which may be existing or under development at IETF.  

Are we prepared to make a stronger statement here?  Are we prepared to
say: 
If you use NAT66, then be prepared for interoperability problems with
IETF specifications because we WILL NOT design around your box, and,
furthermore, that all the reasons you would want such a box have been
fully accomodated through other means which are all in a good enough
state for you to deploy today.

- Wes

-----Original Message-----
From: Iljitsch van Beijnum [mailto:iljitsch@muada.com] 
Sent: Thursday, November 06, 2008 9:15 AM
To: Wes Beebee (wbeebee)
Cc: EricLKlein@softhome.net; Margaret Wasserman; v6ops@ops.ietf.org;
Behave WG
Subject: Re: [BEHAVE] Comments on the NAT66 draft

On 6 nov 2008, at 14:59, Wes Beebee (wbeebee) wrote:

> As we move to IPv6, NAT44, NAT64, and NAT46 will eventually go away.  
> The problem with helping NAT66 (even when that is not your
> intent) is that once it catches on, it'll be in the Internet forever 
> and will never go away.

> "NATs necessary for IPv6, says IETF chair"
> http://www.networkworld.com/news/2008/072109-nat-housley-qna.html

> Once NAT66 gets out, I can imagine even more damaging headlines (which

> conveniently miss all the subtleties of the message in section 3 of 
> http://www.ietf.org/internet-drafts/draft-mrw-behave-nat66-00.txt)
> : "IETF Standardizes IPv6-to-IPv6 NAT".

Well, if that's what we want to avoid, we shouldn't be coy and come out
and say that IPv6 NAT won't be accommodated in IETF protocols.

What seems to be happening today is that we all look the other way and
pretend the issue doesn't exist, because we either assume that of course
there won't be any IPv6 NAT or of course there will. So we are on our
way ending up with the same situation that we encountered with
IPv4: suddenly, it's no longer realistically possible to deploy a
protocol that isn't NAT-friendly, but there are so many different NATs
that it's impossible to be friendly to them all, and many of them
operate is very suboptimal ways that could have been avoided with some
forethought.