[CCAMP] Comments and questions on draft-ietf-ccamp-mw-topo-yang-01

[Stephen Cheng <[e-mail]>]

Dear draft-ietf-ccamp-mw-topo-yang-01 authors,

I have been reviewing the current draft, and I have a number of questions and comments attached below. I am looking forward to your illumination and feedback. Thank you.

Warm regards,
Stephen Cheng


The current draft of ietf-microwave-topology augments ietf-te-topology. I would like to better understand the basis of this design choice.

  *   Is it to enable the modelling of a fully traffic engineered network from L1 all the way to L2/L3/MPLS and beyond?
  *   It appears that many aspects of ietf-te-topology that would unlikely to be applicable to Microwave topology, such as
     *   TTP
     *   TE-Path
     *   Inter-Layer Lock
     *   Transition Link
     *   Can the authors clarify if they would be applicable to a L1 microwave topology?
  *   Appendix A.1 and A.2 examples do not show how any of the attributes defined in ietf-te-topology are used in the context of modelling microwave L1 topology. They only show the attributes defined in ietf-microwave-topology. It would be useful to show such examples.
  *   It would be helpful if a section is added to the I-D to explain which aspects defined in ietf-te-topology would be typically applicable to a L1 microwave topology, and how they should be used in conjunction with ietf-microwave-topology.

I would like to understand how authors intend to model/use connectivity matrix/LLCL in ietf-microwave-topology. I would argue that connectivity matrix/LLCL is of limited applicability for modelling L1 microwave network:

  *   Most microwave-capable devices have microwave links and links of other L1 technology (such as Ethernet and optical).
     *   At a radio site, a microwave device is typically connected to other microwave devices, switches or routers via an Ethernet or optical link
  *   Traffic between different links (whether microwave/non-microwave) within a node are typically switched at layer 2, or routed at layer 3 or IP/MPLS, and not at L1 level.
  *   As such if it is desired to describe traffic engineering between links (whether microwave/non-microwave), it should be described at the layer which the traffic switching/routing occurs.
     *   For example:
        *   a L2 topology (perhaps augmenting ietf-te-topology) could describe the switching capability at the layer 2, with L1 TE topology as the underlay topology
        *   a ietf-te-topology topology could be used to describe the traffic engineering for IP/MPLS, with L1 TE topology as the underlay topology
     *   Due to the fact that a L1 TE topology being an underlay for L2 TE, L3, MPLS TE topologies, the domain controller or operators can still determine if a L1 microwave link is over-subscribed.
  *   If the above reasoning is correct, ietf-microwave-topology should not need to augment connectivity matrix/LLCL with new microwave specific attributes, unless L1 switching is expected on microwaves.

It is a common requirement for users (of a domain controller) to expect a single layer 1 topology to describe all the physical devices and all the L1 links no matter what the L1 technology is. In most networks where microwave plays a part, there would be other L1 technologies such as Ethernet and optical. As such it would be desirable to structure ietf-microwave-topology so that it could co-exist with the modelling of other L1 technologies.

  *   At the moment this is not possible with the proposed ietf-microwave-topology, even although RFC 8345 section 4.1 says
"a network can even have multiple types simultaneously. The type or types are captured underneath the container "network-types""
     *   It is not possible to determine if a LTP is a microwave termination point or another type of link
     *   It is not possible to determine if a link is a microwave link or another type of link
  *   One possible solution is as follow:
     *   Add a "microwave" presence container under nt:tp to indicate that it is a microwave TP
     *   Add a "microwave" presence container under nt:link to indicate that it is a microwave link
     *   the mw-link-attributes in nt:tp are conditional on the "microwave" presence container under nt:tp
     *   the mw-tp-attributes in nt:link are conditional on the "microwave" presence container under nt:link
  *   If other L1 technology topology models follow a similar protocol (i.e. technology-specific presence container for TP and link), it would be possible to have a single L1 TE topology instance to model a network with a mix of L1 technologies.