Re: Going forward with zero-config tunneling requirement

[Pekka Savola <pekkas@netcore.fi>]

Jordi,

I think you were missing one assumption about 3GPP, that their 
solution would only be needed when there is no NAT traversal or 
proto-41 forwarding, i.e., the 3GPP operator provides the tunnel 
endpoint.  If that's not the case, all bets are off.  Below..

On Thu, 30 Sep 2004, JORDI PALET MARTINEZ wrote:
> My view actually is that 2 & 3 could be the same, specially because
> in the real world I've seen GPRS deployment both with no-NAT (public
> addresses) and using NAT (with private addresses). But in the later
> case, NAT traversal was not required because proto-41 forwarding
> worked. Of course, this may be not the case all the time !
> 
> I also understand that having a zero configuration solution w/o NAT
> traversal could be technically very difficult and probably will mean that
> the solution will not be "simple" as required by 3GPP (no new "whatever" in
> the existing infrastructure). Of course, everything needs to be balanced and
> considered.

The assumption of '3GPP' case is that even if private addresses are
used, there is no need for NAT traversal, because the
zero-configuration tunnel server is provided in the 3GPP network,
i.e., in the same 'addressing domain'.

The zero-configuration solution for 3GPP does not try to solve the
case where the 3GPP operator does not provide the tunnel endpoint.  
[Karen: maybe this should be made more explicit in the requirements]

That would be rather close to the 'ISP' zero-config situation.

(FWIW, I think it's certainly a consideration whether a 3GPP UE could
work even when the 3GPP operator does not provide this capability, or
in an enterprise's WLAN environment, or whatever -- in that sense,
having a more generic solution would seem to be rather helpful.  But
that's something that needs to be debated later.)

> Finally, my opinion is that a single document can take both 3GPP and
> non-3GPP zeroconfig, but may be we need to progress a little bit on the
> solution or candidate solutions to assert that ?

Could be (I personally thought so), but there was reluctance to this, 
but we need to make progress in the requirements, and this is one 
approach to accomplish it.

-- 
Pekka Savola                 "You each name yourselves king, yet the
Netcore Oy                    kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings