The MPLS WG Archive

["Krishna Bala" <kbala@tellium.com>]

Tony,
In the extreme cases there need not be any distinction between the
peer model and the open models.

The peer model can clearly support the case where there are well defined
interfaces
to an optical network element and also between optical network elements.
The open model also can support the peer model.

However, in the general case these models are based on fundamentally
different
assumptions:
1. The peer model is based on the assumption that the IP router has full
visibility
into the optical layer. It is able to make wavelength assignment decisions,
it
can route lightpaths based on physical layer impairments (e.g. accumulated
SNR,
PMD, delay, fiber non-linearities ...). I strongly question the basis of
this model.
Yes, I concede that there is nothing that prevents the peer model from
supporting
the open model. But, I am not in agreement with the fundamental premise upon
which
this model is based. I think it is a mistake to assume that the IP router is
going to have such visibility into the optical layer ... at least initially.

2. The open model is based on the assumption that any "client" (e.g. IP
router,
ATM switch, direct wavelength services) can be supported over an optical
layer
that comprises network elements from multiple vendors via the use of well
defined
interfaces. The UNI from the client to the optical layer and the NNI between
the
optical network elements. Again, the open model clearly supports the peer
model also.
I believe that the open model allows us to start early and immediately
with something that gets us going in the right direction without all of
complications involved with the peer model.

Fundamentally I agree with you that each model can support the other. But in
practice
I do not see how we can achieve any interoperability without keeping it
simple
initially.

Krishna


> -----Original Message-----
> From: owner-mpls@UU.NET [mailto:owner-mpls@UU.NET]On Behalf Of Tony Li
> Sent: Saturday, March 25, 2000 4:49 PM
> To: Krishna Bala
> Cc: Jonathan Lang; curtis@avici.com; Jagan Shantigram; Khaled Elsayed;
> mpls@UU.NET; ip-optical@lists.research.bell-labs.com
> Subject: Re: Comments on draft-ip-optical-framework-00.txt
>
>
> > Open refers to the fact that this model supports the interconnection of
> > several types of clients to an optical network (server layer).
> > Examples of Clients:
> > IP Routers, ATM Switches, SONET ADMs, "Direct Wavelength Services"
> >
> > The open model also allows the interconnection of other Optical Network
> > Elements to each other.
> > Examples of other Network Elements:
> > Optical Add/Drop Muxes, Other OXCs, Wavelength Selective XC, Wavelength
> > Interchanging XC
> >
> > In general, since it offers the capability to support ALL
> services (not just
> > IP) it is an Open Model.
>
> The distinction that you make between the Open Model and the Peer
> Model is not
> clear to me.
>
> Certainly both models will support all services.  Both models
> should be using IP
> as the basis of their routing and signalling protocols.
>
> What is the substantive basis for a difference?
>
> Tony
>
>