Re: New charter

[Kireeti Kompella <kireeti@juniper.net>]

Hi Loa,

On Fri, 26 Nov 2004, Loa Andersson wrote:

> there seems to be a majority for doing L1VPN in ccamp, I don't
> particular object, but remember the arguments when we placed the
> L3VPN and L2VPN in the Internet Area.

Thanks for articulating this -- I've been thinking about this as well.

> The main argument was that the Internet Area deals with end-2-end
> services over the Interent, thus VPNs belongs in the Internet Area.
>
> If we decides that L1VPN don't go together with L2/L3VPNs in the
> Internet Area we should at least have some motivation for this.
> What is it?  Is L1VPN not an end-2-end services or is it not an
> Internet Service?

I've come up with four reasons why CCAMP is a better venue for L1VPNs:

a) L1VPNs are definitely not an end-to-end service -- I don't think
   we'll see end-hosts with OC-48s (or even fiber GE) and GMPLS
   signaling any time soon.  However, it is clearly possible for end
   hosts to connect to an L2 or L3VPN.
b) L1VPNs are not necessarily an Internet service (well, for that
   matter, neither are L2VPNs).
c) Most of the work for L1VPNs is done -- starting a new WG for this
   will hinder progress, and also expand work to fill out a whole WG.
   The bulk of the work is documenting what we already have that works.
d) L1VPNs are distinctly different from L2 and L3 VPNs -- in the
   latter, the PEs do all the heavy lifting, and the CEs are then
   immediately connected.  In L1VPNs, the PEs do discovery and
   information exchange, but the *CEs* must do the signaling to set up
   the (currently) desired CE topology.  To do the signaling, the CEs
   must exchange information with the PEs.

This last part requires tight coordination with the CCAMP WG, both for
signaling (GMPLS UNI), and also for information exchange (pretty much
the same as for GMPLS inter-domain).  This is for me the most
important reason why this work should be kept in CCAMP rather than
creating up a new WG.

Kireeti.
-------