Comments on ietf-layer0-types from draft-ietf-ccamp-optical-impairment-topology-yang perspective

[dieterbeller]

This issue is used for collecting comments on https://github.com/ietf-ccamp-wg/draft-ietf-ccamp-layer0-types/blob/8e0d1878b6a0a8e2cc675343175ba269799bf9e6/ietf-layer0-types.yang from an draft-ietf-ccamp-optical-impairment-topology-yang perspective

[dieterbeller]

Dieter's comments (numbers are line numbers):

General comments:

• Is the scope of the very first version of layer0-types spectrum management at layer 0 only?
• layer0-types does not reference ITU-T Recommendation G.807 and does not use the layer 0 media channel concept as defined by the ITU.
• There are many groupings that contain a single leaf - this seems to be a bit contradictory to the purpose of the grouping statement as deifined in RFC6020.

The following comments are based on the assumption that the scope is layer 0 spectrum management only:
54: the topology model distinguishes 3 different types of operational modes - this is an optical transponder property

269: is term-phys the OCh/OTSi termination? G.709 does not use the term "Physical layer termination"

277-299: theses are layer 1 and not layer 0 termination functions - why are they in layer0-types?

301: term-section - is it related to the OTS or OMS?

309ff: layer0-bandwidth-type should be the spectral signal width in GHz and not the data rates of the electrical client signal carried by the optical signal (OCh/OTSi). The OTU rate is a layer 1 property. Only the FEC, which is terminated by the OTU termination function could be considered as a kind of layer 0 property (see fec-type, line 487ff).

511-538: see previous comment - this should be the spectral signal width in GHz.

679: the grouping flexi-grid-node-attributes (plural!) contains just a single attribute: leaf node-type

696: grouping flexi-grid-path-bandwidth is not a spectral width in GHz (see comments above)

710: grouping flexi-grid-link-bandwidth is not a spectral width in GHz (see comments above)

738: grouping flexi-grid-channel - this seems to be a media channel

809-834: are leaf min-slot-width-factor and leaf max-slot-width-factor related to the two multipliers in https://tools.ietf.org/html/rfc8363#section-3.2? Citation:

• "Slot width range: two multipliers of the slot width granularity, each indicating the minimal and maximal slot width supported by a port, respectively."

[haomianzheng]

Dieter's comments (numbers are line numbers):

General comments:

• Is the scope of the very first version of layer0-types spectrum management at layer 0 only?

Yes, as specified in https://mailarchive.ietf.org/arch/msg/ccamp/tSNbefozXaQFjkrRzH7OATAiq5Y/, let's target on preparing the ietf-layer0-types for the first cluster of drafts, i.e., a) b) c) d).

• layer0-types does not reference ITU-T Recommendation G.807 and does not use the layer 0 media channel concept as defined by the ITU.

Is this a general comments or very explicit one? In other words, which identity in ietf-layer0-types do we explicitly need the reference to G.807?
I agree that G.807 is a related document, on the architecture level. However, the current identities in the module are referencing to data plane specifications, instead of architectural concepts. So far I did not see any identity that should referencing to G.807, please suggest if any.

• There are many groupings that contain a single leaf - this seems to be a bit contradictory to the purpose of the grouping statement as deifined in RFC6020.

Could you raise the specific paragraph/sections in RFC6020? I did not find the restriction that "grouping must have multiple leafs". It's nothing but to be efficient as there are multiple 'uses' in the topology/tunnel models.

The following comments are based on the assumption that the scope is layer 0 spectrum management only:
54: the topology model distinguishes 3 different types of operational modes - this is an optical transponder property

Accepted, and will be removed.

269: is term-phys the OCh/OTSi termination? G.709 does not use the term "Physical layer termination"

This needs further clarification. Will remove (move to a new draft as phase 2). Refer to: #5 (comment). They are related to client signal configuration.

277-299: theses are layer 1 and not layer 0 termination functions - why are they in layer0-types?

Same Above.

301: term-section - is it related to the OTS or OMS?

Same Above.

309ff: layer0-bandwidth-type should be the spectral signal width in GHz and not the data rates of the electrical client signal carried by the optical signal (OCh/OTSi). The OTU rate is a layer 1 property. Only the FEC, which is terminated by the OTU termination function could be considered as a kind of layer 0 property (see fec-type, line 487ff).

FEC related identities are same above and will be removed.

511-538: see previous comment - this should be the spectral signal width in GHz.

A general feedback on all the bandwidth-related comments: the bandwidth augmentation for WSON/Flexi-grid is still questionable on how to represent. Options include in a client-signal way (OTU), or bps way, or others. For the sake of spectrum allocation, only using label restriction would be good enough to understand whether the corresponding spectrum is allocated or not.
So the proposal is to move the following to -v2: layer0-bandwidth-type, bw-OTUk..., wson-link-bandwidth, wson-path-bandwidth, flexi-grid-link-bandwidth, flexi-grid-path-bandwidth... No augmentations are needed for both WSON and Flexi-grid topology.

679: the grouping flexi-grid-node-attributes (plural!) contains just a single attribute: leaf node-type

I don't think this is an issue, as there are potential augmentation here. 'attributes' will keep us safe.

696: grouping flexi-grid-path-bandwidth is not a spectral width in GHz (see comments above)

(see comments above on bandwidth)

710: grouping flexi-grid-link-bandwidth is not a spectral width in GHz (see comments above)

(see comments above on bandwidth)

738: grouping flexi-grid-channel - this seems to be a media channel

propose to rename as flexi-grid-frequency-slot.

809-834: are leaf min-slot-width-factor and leaf max-slot-width-factor related to the two multipliers in https://tools.ietf.org/html/rfc8363#section-3.2? Citation:

• "Slot width range: two multipliers of the slot width granularity, each indicating the minimal and maximal slot width supported by a port, respectively."

Yes, will give more description.

[italobusi]

This issue is resolved with 10acf68

I have linked this open issue with PR#3 so it will be automatically closed when the PR#3 is merged