Date
10AM EDT
To Join: https://zoom.us/j/287367106
Discuss
- Introductions
- Review of Action Items
- HPA Modeling
- Discussed possibility of splitting *Requirements attributes into *Requirements and *HpaRequirements
- Discuss VDU vs VNFC Descriptor and relationship to deployment flavors
- Discussed the possibility of extending the model to include deployment flavors at the VNFC/VDU level
- See annex in the ETSI document to be further discussed next week:
https://docbox.etsi.org/ISG/NFV/Open/Drafts/SOL001_TOSCA_desc/NFV-SOL001v040.zip
Discussed need to put a schedule on Papyrus migration
- Recommended Papyrus version: 3.2.0 RC4
Recording:
GMT20180112-145805_VNFSDK-mod_1920x1080.mp4 GMT20180112-145805_VNFSDK-mod.m4a
Upcoming:
Next call 19 JAN 2018
- Identify stable classes in model
- Papyrus port - version, schedule.
- Model driven approach discussion (Alexander Vul)
Model walkthrough
- SDC IM Gap assessment
Discuss VDU vs VNFC Descriptor and relationship to deployment flavors
- VoLTE VNF Descriptor to latest release of IFA011
Gap analysis of HEAT templates
- Event descriptors alignment with DCAE
- New Action items
4 Comments
Chesla Wechsler
Attached is the write-up I did describing just my way of interpreting the ETSI NFV docs for several entity types.
Arun Gupta
In the VNFC v VDU debate: I think for deployment of a VNF, the VDU concept may be sufficient, but for description of service flows through the VNF, one wouldn't want to be talking about VDUs, but rather VNFCs.
Chesla Wechsler
I agree.
Arun Gupta
Regarding
See annex in the ETSI document to be further discussed next week:
https://docbox.etsi.org/ISG/NFV/Open/Drafts/SOL001_TOSCA_desc/NFV-SOL001v040.zip
in my opinion, the connection points of a VNF are either internal xor external. Packets/frames from an internal connection point can only go to other internal connection points. External connection points exchange packets/frames with "outside the VNF". So, contra some notes in the document, IMO, external connection points should not be modeled as references to internal connection points.
PS: are there counter-examples to what I think, where a connection point sends/receives packets both from internal connection points and from "outside the VNF"?