Re: Comments: draft-suryanarayanan-v6ops-zeroconf-reqs-01.txt

[Pekka Savola <pekkas@netcore.fi>]

Hi,

On Thu, 4 Nov 2004, Bound, Jim wrote:
> http://www.v6ops.euro6ix.net/ietf/draft-suryanarayanan-v6ops-zeroconf-re
> qs-01.txt
>
> It seems to be that this zerconf spec, 3gpp zerconf spec, and assisted
> tunneling spec have many common properties and could we reduce this to
> just one spec?

Yes, they certainly share a number of common properties.

However, the problem spaces they're covering seem to be slightly 
different.

In particular, the assisted tunneling assumes explicit user 
registration, which typically maps to providing the service to 
pre-agreed (registered) third parties or as a VPN-like service 
(without encryption).  Registered mode can of course also be used in a 
scenario where registration is not really necessary. If you will, one 
could imagine non-anonymous-mode TSP as an instance of this model.

Now, the explicit goal of both zero-conf documents is to avoid 
registration, and to provide just a simple, very easy way to obtain 
connectivity, without the bells and whistles.  One could consider 
ISATAP run inside an enterprise an example of this kind of solution, 
or a simplified, anonymous-mode, TSP-like solution.

Hence, there is IMHO some value in keeping these separate even though 
there is overlap.  On the other hand, I personally think that the 
documents do not currently describe sufficiently clearly what I wrote 
above, so a reader unfamiliar with the work is unlikely to see much 
difference there at least at first.

That said, if folks think more consolidation is good, to eliminate 
overlap and make it more concise, yes, that's a possibility -- please 
speak up.  This comes with the expense of blurring the (special, and 
conflicting) requirements of different scenarios, though.  Not an easy 
tradeoff.

(We actually tried to start with this project as one spec -- 
draft-ietf-v6ops-assisted-tunneling-requirements-00 intended to cover 
all the cases, but 3GPP folks wanted to separately clearly describe 
their own requirements, and then generalized case of 3GPP was also put 
in a separate document.  So, I guess this argues that there was 
attempt at doing just one document, and it didn't quite work out due 
to differences in requirements esp. from 3GPP..)

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