Last of 4 drafts to OIF : inter-level routing

["Adrian Farrel" <adrian@olddog.co.uk>]

Hi,

A fourth and final draft for you to review...

Thanks,
Adrian

===

Dear Lyndon,

Here is a final response to the issues raised by the OIF. This
one relates to ASON routing and multi-layer networks, and comes
from your communication oif2008.063.02.pdf.

> in our discussions of multilayer networks and impacts 
> on the control plane, we took note of ITU-T Recommendation
> G.7715.1 requirements to advertise:
> a) Local Connection Type; and
> b) Local Client Adaptations Supported.
> We believe that this functionality would be very useful for
> applications that have been discussed in OIF (see
> attachment). We would appreciate further clarification as to
> whether both requirements have been addressed in the current
> CCAMP work on ASON routing support (local connection type in
> particular is not explicitly identified in the current 
> draft). We would also appreciate more information on the
> relationship between support of adaptation in the ASON 
> routing draft and in the multilayer/multiregion networking
> drafts also being developed in the CCAMP WG.

We note that G.7715.1 discusses the operation of multi-layer
networks in an appendix and observes:

  This appendix provides illustrative examples of different 
  methods of carrying multiple layer information between two
  systems. It does not mandate a particular method of 
  implementation for link state routing protocols.

  From the perspective of ITU-T Rec. G.8080/Y.1304, each layer
  network has its own instance of a control plane.

It can also be observed that a single routing level cannot span
multiple network layers within the ASON architecture.

draft-ietf-ccamp-gmpls-ason-routing-ospf-06.txt addresses the 
distribution of routing information between Routing Areas (in 
the ASON sense), but is agnostic to the routing information
passed. For more details of the exchange of "local connection
type" and "local client adaptations" we need to look into the
base GMPLS routing specifications and the CCAMP work on multi-
layer networks.

As pointed out in RFC 4652, the Switching Capability field of 
the Interface Switching Capability Descriptor (see RFC 4202)
describes the capabilities of the receiver with regard to any
specific TE link (e.g., PSC, TDM, etc.).

Your question also opens up the issue of distinguishing 
between what signal types an LSR can terminate (through 
adaptation) as distinct from switch (onto another TE link).
This function is addressed in another CCAMP draft, 
draft-ietf-ccamp-gmpls-mln-extensions-03.txt, which will go to
CCAMP working group last call soon. This draft introduces the
Interface Adaptation Capability Descriptor (IACD) to carry the
necessary information.

Both the ISCD and the IACD can be exchanged between routing 
levels using the techniques described in draft-ietf-ccamp-
gmpls-ason-routing-ospf-06.txt. You may, however, also like to
look at the appropriateness of the PCE architecture (RFC 4655)
in the role of inter-level routing as it seems to fit well with
the architecture as described by the ITU-T in G.7715.2.

Best regards,
Adrian Farrel and Deborah Brungard
IETF CCAMP Working Group Co-Chairs