Re: WG split: why and how?

Folks-

Regarding the dead-line for this discussion and making the decision.
Given the activity on the list, we'd like to give it till the end of
this week. We _will_ make the decision then.

Addressing what hasn't been yet:

Sunday, May 11, 2003, 8:59:57 PM, Don Fedyk wrote:
> ...I suggest it would be more fruitful to spend time
> from now to Vienna advancing current work and seeing what is holding up 
> the rest.  Lets get this train back on the rails before, if necessary, 
> we split the tracks.

From the perspective of advancing the current work, there's no delay
associated with this discussion. For example, L3 requirements and
L3 framework are on the IESG agenda for the telechat this week.

Monday, May 12, 2003, 2:52:14 AM, [email protected] wrote:
...
>> Could you send the draft names, pls?

> Note that I'm not implying that they should be in the first charters. I
> understand your 'basic things first' reasoning. Just wanted to say that
> they've had little attention. A few examples:

I got them on my list for further discussion if we decide to go
with the split. Thanks.

Monday, May 12, 2003, 10:59:37 AM, Yakov Rekhter wrote:
> Agreed. In fact, if the justification for splitting the PPVNP WG
> into two is to allow discussion for L3 services, then following
> the same line of reasoning there should be not one, but two L2 VPN
> WGs, one for L2 VPN services (other than VPLS), and one for VPLS.
> Yet, the IESG said nothing about having two L2 VPN WGs. Perhaps the 
> IESG would be kind enough to explain this inconsistency.

Thanks for trying to catch a possible inconsistency.
"Following the same line" can take one quite far in splitting big
WGs in smaller ones. One clearly has to stop somewhere. Based on
the conversations we had, splitting between L2 and L3 is the most
reasonable choice.

Monday, May 12, 2003, 12:33:08 PM, Ash, Gerald R (Jerry), ALABS wrote:
...
> Seems to conclude that the problem cannot be solved within the
> existing PPVPN WG. Let's get the WG chairs' proposals on how to fix
> the problem within the WG, and discuss their proposals. Let's reach
> a WG consensus as to which solution best solves the problem.

To repeat what I stated already--WG management decisions are not made
by consensus within the WG. It is the AD's and the IESG
responsibility. The ADs will consider the discussion on the list
and conversations with the WG chairs and will make the decision.


Monday, May 12, 2003, 1:20:26 PM, Hamid Ould-Brahim wrote:
...
> If we assume a wg split is necessary (which I am no sure it is
> at this point in time), then why not address VPLS 
> separately from other l2vpn work. A potential option is Yakov's suggestion 
> to split l2vpn in two wgs.

See my answer to Yakov above.

> Another option is to:

> a) Keep ppvpn wg as it is covering l3vpn and l2vpn
>    (other than VPLS), still covering the generic documents, 
>    the common documents/mechanisms, the "Provider-Provisioned" 
>    aspects, inter-as, etc.

I think this would still be too much for one WG.

> and,

> b) Move VPLS specific related work (including provider-bridges liaison,
>    etc) in a "Temporary" focused wg (we have temporary area why not
>    have "temporary WGs" in INT area as well). 

FYI, formally all WGs in the IETF are considered "temporary", leaving
only as long as their charters allow. When all milestones are
completed, the WG potential is reconsidered.

>    While the ppvpn, vpls are progressing we can decide to join them 
>    at a given time. 

> At least this option keeps ppvpn as it is in terms of structure,
> avoids the questions on what to do with common l2+l3
> work, does minimum disruption to ppvpn wg activities in l3, l3+l2 
> space, takes into account that l3vpn work is in its final stage.

I suggest you look at it this way: L3 WG would be a rechartered and a
focused version of PPVPN with L2-related work removed. The discussion
about common L2/3 documents is yet to happen, and if we decide to go
with the split, this will be one of the first things to discuss. We
got the message about importance of these items, rest assured. Wrt
l3vpn in its final stage--it's not going to be affected.

> Other options as well is to look at a re-org from a wider
> angle that affects both PWE3 and PPVPN. For example, merge some
> PWE3 work with l2vpn VPW work, etc (I never understood why 
> some l2vpn signaling is done in another wg).

Seems to me PWE3 has enough workload... Thanks for the suggestion
though, we'll discuss it.

> It looks to me an "easy-step" approach is more appropriate than
> drastic change where the vpn problem is simplified into 
> l3 and l2 work.

I don't believe anyone is thinking that L2 and L3 things are perfectly
splittable. Just like between RTG WGs or in the MPLS/CCAMP case, there
are a lot of connections. We are well aware of that and will deal with
this. Splitting the work in two WGs is targeted to help manage the
workload.

> I definitely favor (as Jerry already noted) a 'let's wait a bit longer'
> approach and progress the current work and if needed discuss in Vienna 
> options and solutions as was suggested by others (and 
> involving all the parties). 

I think I replied to a similar proposal before--uncertainty wrt the
question under discussion is not a good thing. We (the ADs) need to
make the decision sooner than that.

Alex