Comments on layer0-node-type

[italobusi]

I have few doubts on the layer0-node-type identity, which defines four types of nodes: flexi-grid, WSON-FOADM. WSON-ROAM and WSON-iLA

1. I assume that a flexi-grid node is a flexi-grid ROADM (FOADM can be considered old type of equipment not supporting flexi-grid). Correct?
2. I am not sure why iLA is only defined for WSON. I think it is also possible to have iLA used within flexi-grid topology.
3. Considering the previous comment, I am also not sure we need different identities for WSON and flexi-grid nodes, since this information can be inferred from the topology type (WSON or flexi-grid). Therefore, I am wondering whether the node type should not just define three types (ROADM, FOADM and iLA), maybe with some constraints about the fact that FOADM can only be used by WSON topology
4. I am not sure the ITU-T references are correct: G.694.1 defines the DWDM grids and not the layer0 node type nor the ROAM nor the FOADM. G.698.2 defines the application codes for black link interfaces and not the ILA
5. What about the case where a TE Node is an abstraction of multiple ROADMs?

[haomianzheng]

I have few doubts on the layer0-node-type identity, which defines four types of nodes: flexi-grid, WSON-FOADM. WSON-ROAM and WSON-iLA

1. I assume that a flexi-grid node is a flexi-grid ROADM (FOADM can be considered old type of equipment not supporting flexi-grid). Correct?

Yes.

2. I am not sure why iLA is only defined for WSON. I think it is also possible to have iLA used within flexi-grid topology.

Right. Regarding the spectrum allocation (phase 1 of ietf-layer0-types), the iLA may not be needed. Then shall we remove the iLA in the current phase?

3. Considering the previous comment, I am also not sure we need different identities for WSON and flexi-grid nodes, since this information can be inferred from the topology type (WSON or flexi-grid). Therefore, I am wondering whether the node type should not just define three types (ROADM, FOADM and iLA), maybe with some constraints about the fact that FOADM can only be used by WSON topology

Propose to be just ROADM/FOADM, the key is 'whether reconfigurable or not'.

4. I am not sure the ITU-T references are correct: G.694.1 defines the DWDM grids and not the layer0 node type nor the ROAM nor the FOADM. G.698.2 defines the application codes for black link interfaces and not the ILA

Well, G.694.1 defines the grid, but not the node. The node is defined somehow in G.680 (as optical network element) or in G.671 (OADM as a subsystem). It is proposed to change the layer0-node-type(to be l0-node-type)/wson-node-foadm/wson-node-roadm to G.671. The reference of identity flexi-grid-node is correct, should be G.694.1.

5. What about the case where a TE Node is an abstraction of multiple ROADMs?

This is possible, and should be an implementation issue. It is possible that a TE node can either be one or multiple ROADM/FOADMs. The controller should be capable to locate the ROADM/FOADM(s) if there is any abstraction.
From the standard perspective, any single TE node above the PNC (on either MPI or CMI or MMI) should be in an abstract format.

[haomianzheng]

Based on offline discussion, the conclusion is:
Use presence container 'wson-node' and 'flexi-grid-node' to differentiate the WSON with flexi-grid. Then indicate whether the WSON node is reconfigurable (ROADM) or not (FOADM). The proposal is to have a 'is-reconfigurable-node' only applicable for wson-node in ietf-wson-topology (not in ietf-layer0-types). This is captured in ietf-ccamp-wg/draft-ietf-ccamp-wson-yang#1.

For ietf-layer0-types, remove all the 5 identities, layer0-node-type, flexi-grid-node, wson-node-foadm, wson-node-roadm, wson-node-ila.

[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