RE: Recent navel gazing - we need to stop wasting cycles on FUD

["Tony Hain" <alh-ietf@tndh.net>]

Pekka Savola wrote:
> On Mon, 22 Mar 2004, Alain Durand wrote:
> > > I don't think any serious enterprise would want to do anything as
> > > unreliable as 6to4.  Just get a configured tunnel w/ prefix delegation
> > > or native access.
> >
> > I'd like to think of the company I work for as being serious.
> > We are using 6to4. Not in the way you think, but we are still using it
> > very seriously.
> 
> It's an entirely different ballgame if you don't use 6to4 for external
> connectivity (but internal only, as I recall from previous exchanges)
> .. so for the general 6to4 commectivity, I believe the comment is
> still rather accurate.

The point is the validity of the statement really doesn't matter. There is a
valid use for 6to4, as there is with all the other transition technologies
on the table. There is no duplication between them we can identify (other
than possibly in the tunnel broker collection), so we appear to have a
minimum set. Thus they need to be standardized, which may include
modifications from the current documents. 

We also have a set of environment descriptions, and a charter to identify
which of the standardized technologies applies. Specifically 6to4 is on the
list of Enterprise focused technologies, so how it gets used is what needs
to be documented. We do not need to waste time trying to kill any one of the
technologies simply because it does not provide a one-size-fits-all answer,
or does not fit someone's NIH profile. While manually configured tunnels may
appear more stable, they will only scale to a point. For the period beyond
the scale limitations of manual tunnels and full deployment of native
services, an automated tool like 6to4 allows deployment to continue. If any
one network manager chooses not to deploy it, that is a local choice. The
job of the IETF is to get the standardization work done, then provide hints
about intended uses and potential challenges. The market is perfectly
capable of figuring out which tools really apply to the local situation, and
nothing the IETF publishes will change that. If the IETF chooses not to
standardize the technologies or provide usage hints, the market will be left
to pick from non-interoperating approaches, but will none the less meet its
needs to achieve a haphazard deployment. 

Tony