Assume that specification of hardware platform capability requirements for R2 will be based on TOSCA templates. TOSCA templates may optionally contain capability clauses defining hardware platform capability requirements. The capability requirements will be specified as comma separated lists of tuples. The ONAP information model specifies the following sets of (optional) hardware capability requirements - CPU, memory, storage and network. The requirement sets are associated with specific VDUs. If hardware capability requirements are provided, SDC will need to parse them and internalize them in a format suitable for use during initial instantiation and operation. Tuple parsing tools will be required to parse requirement strings. We can use the same tools as the ones provided for VNFD validation, as part of the VNF SDK, for this purpose.
Questions:
- What is the SDC-internal VNFD format look like? JSON? YAML?
- Is the VNFD contained in the catalog, or elsewhere?
- How does SO extract the VNFD information from the SDC?
Hi All,
Here is the summary and actions from today’s call:
- Support for HPA assumes TOSCA-based VNF
- Support HPA for HEAT-based VNF is currently out of scope. If it requires, additional discussions to discuss if and how to support it in R2.
- HPA information should be captured per VDU/VNFC. Anatoly to review and come with proposed modeling approach.
- Michael to schedule a meeting to discuss VID/SO “orchestration type” solution.
- Michael/Lior to schedule a follow-up discussion with Alex to decide what work Intel’s developers can support.
Optional scope:
- HEAT to TOSCA and license model builder
- Onboarding - Image upload to Multi-VIM
- Onboarding – SDC/VNFSDK integration
- Implement SDC model
- Michael/Roy enter the relevant epics to SDC and VID