RE: Strategy for VPN work in IETF

Alex et al,

Having observed the discussion this week, I certainly agree 
with the many others who are concerned about the justification 
for the WG split and the manner in which the restructuring is 
being proposed.

Many issues are being mixed together, including ways to speed 
up the work, rechartering, moving the work out of the sub-ip 
area, and the WG split itself. Splitting the group is not a
pre-requisite for addressing the first three issues, and I 
agree with those who argue that a convincing case has yet to 
be made that having a single WG is a real problem.

The proposed charters seem to me to be a step backwards since
they have removed any reference to common components and 
mechanisms which was a goal of the current charter. Issues
like VPN discovery, membership and topology are very similar
whether data-plane forwarding takes place at layer-2 or  
layer-3, and I think that development of common mechanisms is 
something to be encouraged. For example, as was pointed out 
before, Juha's VPN discovery draft using Radius, and Hamid's VPN
discovery draft using BGP, seem equally applicable to layers
2 and 3, and splitting the group would make the development of
common solutions more cumbersome, and less likely. 

As regards speeding up the work, ideas such as a second session
at IETF meetings, interim meetings, forcing resolution of some 
controversial issues, etc, all warrant examination, and I'm sure
that those more closely involved in the work have more ideas
on this topic; however I doubt whether splitting the group will 
by itself make much difference. 

A rechartering exercise may be useful at this time if it ensures 
that later progression of documents through the IESG will be not 
be subject to undue delay (for example if the charter calls for
standardizing multiple solutions, then the fact that multiple 
solutions were produced is not a valid objection to progressing 
those solutions). Also I've never viewed a PPVPN as a sub-ip 
phenomenon, so moving the work to the Internet area makes sense 
to me.

Lastly I think the new procedure for extending protocols is 
still too complex (e.g. I find the proposal that the VPN 
framework document be updated for every extension very odd).
Most protocols these days are designed to be easily extensible.
There is merit to having wide review of new solutions, however
having a formal process that requires the AAA WG to approve of and 
do a synchronized last call for something as simple as a Radius 
extension, for example, is very cumbersome, and will certainly slow 
things down.

Regards,
Bryan