Re: RES: RES: ReN: IPv6 tranisition issues

On Tue, 7 Jan 2003 08:21:09 -0200
"Marcelo Barbosa Lima" <[email protected]> wrote:

>=20
>   Folks,
>=20
>    I think that Keith=B4s answer did not offend to me. I want only a go=
od
> debate about IPv6=B4s subjects. I really believe that is possible to wr=
ite code
> STATEFUL to NAT where whatever application can be supported. I believe =
that=20
> all connections/sessions related to one application can be inpesctioned=
 (in=20
> application layer) and port numbers and IP address translated like in=20
> simple FTP. =20

you are still incorrect.  I am sorry if this sounds impolite or offensive=
, but=20
facts are facts.  Even 3-party FTP fails in the presence of NAT.

think of it this way - if there are more than two hosts involved in an=20
application, then it's not necessarily the case that a single NAT can see=
 all=20
of the addresses that get passed around between hosts.  if a NAT is betwe=
en
host A and host B, there may be ways that host A and host B can learn=20
addresses for each other without the traffic passing through that NAT.

if you want to discuss this further, please do so in private mail.  it's =
not
relevant to IPv6, since NATs have no place interfering with communication=
=20
between IPv6 hosts.

Keith