ONAP R9 Resource IM Call 2021-6-21

• Date and Time: 2021, June 21, 9pm~10pm Beijing Time, 9am~10am US Eastern

• Meeting Room: https://zoom.us/j/97595070639

• Meeting Recording:  

• Meeting Chat Log:  

Agenda:

• Agenda bashing - 5 minutes

• Model Proposals
  
  

  • ETSI CNF support - ? minutes

  • CNF runtime model - 20 minutes

  • PNF instance model - ? minutes

  • CPS model - 0 minutes

  • Topology model - 0 minutes

• Modeling Documentation - 0 minutes

• Model proposals:
  
  

  • Abstract Topology Model

• Modeling requirements:

  • https://lf-onap.atlassian.net/wiki/display/DW/ONAP+R9+Modeling+High+Level+Requirements

• Related information:

  • Allotted Resource for Network Slicing

  • R8 5G Service Modeling

• Clean/Approved models:

  • ONAP Information Model - Clean and approved models across ONAP releases

• Model Specification:

  • Honolulu release: https://docs.onap.org/projects/onap-modeling-modelspec/en/honolulu/ONAP%20Model%20Spec/im/index.html

• Agenda bashing

  • reminder for filling R9 high level requirements: https://lf-onap.atlassian.net/wiki/display/DW/ONAP+R9+Modeling+High+Level+Requirements

• Model Proposals
  
  

  • ETSI support for CNF model

    • R9 DM proposal: https://lf-onap.atlassian.net/wiki/pages/viewpage.action?pageId=16463007

      • aligned with SOL001 v4.2.1, package format is still under discussion

      • will ask for approval after SOL001 v4.2.1 is stable

    • a proposal of updating the information model documentation is planned in R9

    • latest updates:

      • Thinh has presented the alternative package solution in the CNF taskforce call, could be discussed in future calls

      • CNF requirements collaboration pending discussions as well, could be discussed in this week's CNF taskforce call

    • plan to have a conclusion next Monday, and then share to modeling team

      • current gap is that whether NFVO needs to understand the helm chart, and whether ASD can be totally non-overlap with helm chart

  • CNF runtime model

    • Proposal:

      • for Istanbul: add some necessary information as an enhancement of vf-module, to allow components to query k8s objects

      • for Jarkata: define a cnf-module, to store information like compute (pod), storage, etc.

      • the two sub-proposals can co-exist

    • Update:

      • Open

        

      • the idea is to capture the basic information from k8s

      • the proposal has been presented to the A&AI team, and no major concerns

    • Questions:

      • if A&AI team can support modification of existing models (in case in later release, ETSI alignment team/ETSI standard identified changes needed for the runtime model), Lukasz will help check with A&AI PTL

        • William's respond is the update is possible, but A&AI team would like to avoid data migrations; their suggestions is to define the model as generic as possible to prevent massive changes in the future

        • Jacqueline's feeling is that adding attributes in A&AI is simpler than changing the existing model

      • align with ETSI alignment team

    • Conclusion:

      • will ask for poll on the data model

      • will continue how to document the information model next week

  • 5G service / CPS model

  • PNF instance model

    • reverse engineering
      
      

      • feedback to A&AI team:

        • some relationship lack description, and the existing description may not be meaningful enough

      • current status: AAI Schema Reverse-Engineering to Papyrus

  • Topology model

    • Proposal: https://lf-onap.atlassian.net/wiki/display/DW/Abstract+Topology+Model

      • Martin will update the models with the IM sketch

    • Model Sketch: https://wiki.onap.org/display/DW/Proposed+Topology+IM+Sketch
      
      

      • inspired by TAPI model to add "link", "connection" classes

      • looking at ORAN use cases and see if the models can fulfill the requirements

      • TAPI presentation may be found here: otcc2019.KS.002_TAPI_Concepts_Topology_Connectivity.pptx

    • Andy has helped to create a high level requirement for the topology model

      • expecting ORAN to provide relavent use cases

      • Martin and Andy are aware of the different views of topology model presented last week by Cagatay

      • mainly triggered by ORAN WG10, also related to WG1's slicing use case, etc.

      • how to model different layers, physical, logical topology, etc.? current methodology is using UML/papyrus to capture the models

    • Latest updates:

      • Andy presented the updated diagram (names are better aligned with the TAPI)

      • plan to have a presentation to the ORAN team

• Modeling Documentation
  
  

  • gendoc template discussion:

    • whether the description only applies for a navigatable end of an association

    • add relationship documentation to the UML guideline: ONAPMODEL-28: Using the Documentation Property for RelationshipsOpen

    • UML principle wiki page:

      • Principle and Guideline for Information Model Proposal Using Papyrus (Draft)

      • review the updates to the Gendoc requirements part

        • Michela suggest whether it's possible to add information of associations into the class definitions

          • Xu will investigate it offline

      • Using Papyrus for Modeling updates