Uploaded image for project: 'ONAP Architecture Committee'
  1. ONAP Architecture Committee
  2. ONAPARC-269

Architecture and Modelling assumptions for NSD (internal to ONAP)

XMLWordPrintable

      2019-04-01 F2F Subcommiittee

      Thinh presented a summary of the work so far (link to presentation): https://wiki.onap.org/download/attachments/53248195/ONAP%20NSD%20discussion%20v1_final.pptx?version=1&modificationDate=1554150250016&api=v2 

      • The presentation had two parts - analysis summary and recomendations
      • No questions on the analysis
      • Discussion in the summary
        • Nested Services was commented, but it introduced into SDC.
        • Comment to stop focussing on node types, and instead use other types.
        • Comment: if having a proper tosca orchestrator, hence the having the same type definitions should not be an issue, so if this is an issue then it represents that ONAP is not leveraging tosca orchestration to its full extent
        • There was a comment that it would be helpful if the listed issues could relate to what it means to ONAP.
        • There was a question which goes beyond this which was how to handle the different VNF packages.
        • There was a request to sit down with SDC, SO, Multivim on the implications
        • There was a disucssion on that we have one internal model as the VFC/Usecase UI approach is another path

       

      There are two potential representations of the network service:

      • Option A: Implicitly represent the network service
      • Option B: Specifically represent the network service

      After the January Developers event, the identified way forward was:

      • Describe how the flows for onboarding and distribution.
      • Describe the model representation
        • Static "definition" view
        • Dynamic "run-time view" in AAI
      • Describe the flows for instantiation

      We need to discuss the template, whcih is represented at: https://wiki.onap.org/display/DW/Internal+NSD+representation 

       ------------------------------

      2019-02-12Thing shared his findings on the ETSI NSD and the service generated by SDC. https://wiki.onap.org/download/attachments/50202249/ONAP%20NSD%20discussion%20v0_7_wo_conclusion_recommendation.pptx?api=v2 This contained the current findings, and the proposed conclusins are still be worked on.

       

      -

       

       

      2019-01-29


      Gill started with the presentation the implicit representation.

      • It was clarified that there is green text to discuss the variants of passing the application data to the VNFs, however that doesn't influence the implicit or explicit representations.
      • The question was raised whether was requirements were common
        • It was commented that the inner service or nested service concept could be seperated.
        • Requested to solve this for one domain, one service, for now.
        • Gil responded that he introduced the inner service (nested) service is limited in the sense that ONAP doesn't know whether there is some extra out-of-band configuration ongoing.  Therefore ONAP doesn't need to understand whether ONAP is managing the Network service or the full E2E service, it just knows that it knows.  This is the main point behind the proposal.
        • The discussion lead to
      • There was a comment regarding the NSD onboard and converting into an internal model may have some challenges.
        • It was rebutted that the service model is not static and can be enhanced.
      • Thinh indicated that in ETSI they use the term "virtulized" and "non virtulized" data.  Gill pointed out that the NSD is well defined and we can work with that for now.
      • For the scenario where ONAP exports a NSD to the NFVO
        • There is a discussion whether its needed to convert the NSD or send the original NSD.
      • Maybe there is a need to model the externally managed service that is different to an internally managed service
      • Need to think about when exporting a NSD whether there is a need to export the onboarded NSD or whether it has to be generated dynamically.
      • Q: Can we evolve the AID model to be more compatibel to ETSI-NSD, but still be able to be extended.
      • Wait for the analysis of the SDC on the service model. (Thinh)

       

      2019-01-15


       

      Gil

      Description:

      Slides

      There has been a dicussion about whether the NSD is seperate or the same as the e2e service.   It is based on allowing the operator to have different options and support ETSI with a logical translator at the edge.

      2018-10-30 Arc dublin F2F:

       Gil presented.

      There were questions regarding whether there is acomparnison between network service instance model adn the service instance model in onap

       

            auztizza auztizza
            auztizza auztizza
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

              Created:
              Updated:
              Resolved: