Re: IMPORTANT: Strategy for VPN work in IETF

Hi, Paul,

> The proposed split seems to put very little emphasis on the "PP" of PPVPN.

> The proposed focus on technologies for L2 and L3 is likely to seriously
> disrupt the focus on the "Provider Provisioned" concept, where much of the
> value comes from aiming for interoperation across all major VPN
> technologies.  Of course, we could have PPL2VPN and PPL3VPN, but as has been
> pointed out, we then need to coordinate the "PP" parts across the two WGs.  

I think before we go and talk about interoperation across all major
VPN technologies, we should first make visible progress on those
technologies, which focused technology specific WGs are aimed to help.

Regarding coordination--totally agree, we will indeed need to
coordinate, we actually have an idea for this--create a PPVPN
directorate that will ensure such close coordination.

> VPN customers want a bundle of VPN services, and don't really care whether
> it is L2 or L3 (or L1) or some mixture.  VPN service providers want an
> interoperable toolkit which they can use to design services that they can
> differentiate in various ways to attract VPN customers. PPVPN is providing a
> forum to define the operation of the VPN tools.  Regardless of L2 or L3
> technologies, there is a need for simple and interoperable mechanisms for
> VPN membership and discovery, security considerations, SLA verification, and
> a host of management areas.  These mechanisms are being defined, clarified,
> and debated in PPVPN mostly without regard to the specific L2 or L3
> approaches.  They should not be either dropped, or split across two WGs.

I don't think anyone suggests to drop those. First, we are not
splitting the community, only the WGs as organizational entities to
help scaling the management part of the process. People should still
talk to each other and discuss those. If folks believe something
common for both WGs needs to be worked on, we'll put it in one of the
WGs and will inform the other. Occasionally we can even have joint
meetings, which happens in the IETF.

It seems to me that at this moment, technology-specific work items
should be high priority. Most of those pieces are not through the
system yet.

> Yes there is a lot of work in the group, but if it gets fragmented and
> things slip through the cracks, then we will be forcing the VPN service
> providers, vendors, and customers to do much more work and spend much more
> time in the long run.

> In terms of meeting logistics, we will find that many of the same people
> will be attending and developing drafts for the proposed L2 and L3 WGs, so
> we will have to make sure that they don't overlap... Just the same as if we
> plan extended meeting times for the current PPVPN when it is needed.

Same people attending different WGs is perfectly fine, this is a
normal situation in the IETF (examples: RTG WGs, CCAMP and MPLS).
Scheduling-wise, +1 WG is not going to make a big difference.

> Finally, I'm closely involved with two areas of work in L3 VPNs in PPVPN,
> and we are getting to the point where much of the basic work is stable, so
> the L3VPN may not justify a dedicated WG for much longer, although it does
> need a WG in which to operate.

Lack of "heated" discussions does not mean everything is done. There
are items that just need consistent attention and concentration of the
participants to be finished, such as security or MIBs. There is also
more stuff coming in the L3 space.

> I would suggest planning for a longer meeting in Vienna (if the agenda
> justifies it) and bringing this proposal up for discussion there.

Noted.

Alex