Hi Christian,

Christian Huitema wrote:
>> From: Eric Klein, Sunday, February 01, 2009 7:38 AM
>> 
>> The original threads that lead to this list being created all came
>> down to the basic Boolean question: Do we in the IETF want to allow
>> NAT66 at all in IPv6 even though IPv6 was originally designed not
>> to allow it?
> 
> Reality check. The IETF has no protocol police, let alone network
> police. The IETF cannot allow or disallow things, at least not
> through an IETF pronouncement. The IETF only has three choices:

That's correct.

> 1) The Pilatus option: the IETF officially says that it does not like
> NAT, and washes its hand from the outcome.
> 
> 2) The Attila-the-Hun option: the IETF dislikes NAT so much that it
> starts engineering protocols that absolutely break if someone dare
> deploy NAT.
> 
> 3) The Munich option: the IETF recognizes that the evil won't be
> stopped, so dirty its hands and tries a compromise.

My problems with option 3) are:
* people will start to use NAT66 (once available), because they
  are just used to have a NAT for IPv4 (and probably additionally
  confuse it as a security measure). This effect should not be
  underestimated, no matter what health warnings the IETF puts
  in RFCs, which are probably read by the implementers, but not
  by the users/"operators" then. If there is no NAT for IPv6,
  people need to think about why this may be case and look for
  other solutions.

* even though NAT66 does some clever tricks to cause less pain
  for the address translation process itself, the problem with
  IP addresses somewhere in the payload remain. Thus, NAT66 will
  be an obstacle for future applications (otherwise every application
  must be made NAT-aware, use STUN etc.)

Regards,
 Roland
_______________________________________________
nat66 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nat66

Reply via email to