1. Pages
2. …
3. Open Transport Configuration and Control Home
4. TAPI
5. TAPI Call-in Details and Notes

2023-05-30 TAPI Meeting notes

All

Preliminary agenda:

• Review the Feature priority for TAPI "2.5"
  • Hypothesis to postpone Path Computation to a future delivery
• Progresses in gNMI streaming and delta streaming
• Continue the Path Computation UCs analysis
• Continue the discussion of adding OMS CEP spec to version 2.1.3
• Model of the internal connectivity matrix of a ROADM

Brief discussion on optical impairments as routing constraints.

• Agreed that optical impairments can be specified as routing constraints for both Path Computation and Connectivity Services.
• Jonathan Sadler OSNR and Chromatic Dispersion are parameters which summarize the service quality, hence could be candidate as routing constraints.
  • Esther adds that Differential Group Delay gives similar information of Chromatic Dispersion.
• The team agrees that the reference remains A YANG Data Model for Optical Impairment-aware Topology

Nigel Davis presents text specifying delta streaming, which will be included in future version of TR-548.

• Presented a number of principles dealing with changes, clarifying (already available) YANG structures.
• Graphics explaining behaviors required at the Provider.
• Andrea Mazzini suggests to add, where appropriate, the reference to existing tools like Kafka, gNMI, e.g. in case where a combination of more abstract parameter values ends up simulating the default behavior of a specific technology. Agreed.
• Log compaction clarification.
• The change of a record can deal with a complex change by reporting an entire tree, for example the addition/removal of items to/from a list.
• It may be useful a temporary adjustment of the dimension of the "no compaction" log, e.g. when the operator knows that the disconnection is short/temporary, to prevent more costly realignments.
• Work in progress for Protobuf mappings.
• Gabriele Galimberti asks whether it is possible to disable the realignment.
  • Nigel Davis yes through the baseline attribute.
  • Mentioned also the sample qualifier, which currently defined values are partial, missing, invalid, suspect.

Andrea Mazzini shows slides 179-181 from updated otcc2022.AM.001_TAPI_RIASlides.pptx

• Definitions and Use Cases are preliminary agreed:

• Slide 181 raises the question: May a “compute-only” Connectivity Service (including full route info, even physical?) replace the Path Computation Service?
  •  Nigel Davis it seems quite disrupting the complete removal of Path Computation feature.
  • Andrea Mazzini it also seems true that there are several potential overlaps.
    • For the forwarding capability exploration the Connectivity Service may be enough.
    • The Path as a set of links may not satisfy such forwarding capability exploration.
    • On the other side, overloading Connectivity Service (and Connection, CEP, Route) may result in complex/confusing model.
• Showed also slides 189, 190 for multi-ended Path Computation Service.
• Finally agreed that further discussion is necessary regarding requirements and Use Cases.
  • Agreed to re-check the Feature priority for TAPI "2.5"