RE: L3 VPN Work Item List ??

Just realized I did not explicitly include Eric's work on 2547bis with OSPF
as the  PE/CE protocol, but with his usual careful work already proceeding
on it, it should not add a lot to the proposed WG load.

- Paul
-----Original Message-----
From: Knight, Paul [BL60:1A00:EXCH] 
Sent: Friday, May 09, 2003 11:22 AM
To: Thomas Narten
Cc: Alex Zinin; [email protected]; Wijnen, Bert (Bert); Peterson,
Jon; Allison Mankin
Subject: L3 VPN Work Item List ??


Hi Thomas, Alex and all, 

I'm still concerned about splitting off L3 work when it already is mostly
winding down. 

Being involved as editor or co-author of drafts in two of the three L3VPN
types (VR and CE-based) as well as the security framework draft, I don't see
any indication that the work is getting squeezed out of the WG meetings,
only that it is proceeding at the pace of the authors' level of input.  And
it is proceeding  well, if we look at the specifics.

On Wednesday, Thomas Narten said:  "My understanding after having looked at
the work as that there are something like 12-13 IDs on L3 VPNs in the
pipeline. That's no small number and suggests that an L3 focused WG still
has plenty on its plate."  (And this was repeated twice more in the message
under the thread "RE: IMPORTANT: Strategy for VPN work in IETF".)

However, when I look at the work for the proposed L3VPN WG, it seems that
there are four main areas: 
- 2547bis   (5 drafts at least) 
- Virtual Router (VR)  (3 drafts) 
- CE-based IPsec  (1 now, will be 3, with A.S. and MIB) 
- CE route authentication (1 draft) 

Of the four main areas, two are well advanced: 
- 2547bis and its applicability statement have already moved to last call
(2 drafts) 
- VR and its applicability statement have moved to last call.  (2 drafts) 

The MIBs for each of these are basically stable and are typically not
controversial (2 more drafts), and I think also the drafts defining IPsec
and GRE for PE-PE encapsulation in 2547bis are pretty much done, so they
should follow like dominos. (2 more drafts)

Some items are listed in the L3VPN draft charter which really cover both L2
and L3, as far as I can see: 
- Security Framework 
- Textual Conventions (TC) MIB. 
- Also, (but not on the draft charter) Juha's RADIUS-based PE Discovery
seems to apply well to both L2 and L3, and Hamid's BGP Autodiscovery is
applicable to both L2 and L3.

That really only leaves the CE-based IPsec (which will have a MIB and app
statement drafts) and CE route authentication as identified work items which
apply only to L3.  The CE-based IPsec has a strong start and could become a
WG draft in a short time. The related Applicability Statement is also on the
way and will follow.  So I count about four out of the thirteen drafts with
some work.  

This hardly seems like enough work for a full fledged work group.  As Yakov
and others have pointed out, there has been little L3 discussion on the
list... Not because it is being squeezed out, but because it is pretty
stable and has essentially been in the queue waiting for the L3 requirements
and L3 framework documents as stable references.

I'm probably missing something in my quick count here, but it looks really
sparse. 

Paul Knight 
"The world is a jungle in general, and the networking game contributes many
animals."  - RFC 826