Questions: https://lists.onap.org/g/onap-cnf-taskforce/message/209

  1. What format/language is the descriptor written in?
    1. Simple YAML?
    2. TOSCA?
    3. YANG?
    4. Other?
  2. Does the NFVO need to read and understand the Helm charts?
    1. If so, is a full interpreter required?
    2. How much parameterization needs to get resolved?
  3. Where are the required resources (CPU, Memory, Storage, ...) for each Pod described?
    1. Namespace quotas?
    2. Required (initial deployment) or Limit (runtime/scaling) Quota
  4. Where/How are the container images described?
  5. Where/How are the uServices described?
  1. What format/language is the descriptor written in?

[Note] ASD information model is flexible enough to be able to support TOSCA, YANG, or other language.  In the case if NSD is TOSCA language, it makes sense the ASD would be TOSCA.

Note: No Determination.  Can be any of the languages.  Potentially may have criteria from NSD.   Topic for stage 3, protocol discussion.  In general, I may want to move to modeling call. 

  1. Does the NFVO need to read and understand the Helm charts?
    [Note] Yes. ASD depends on the Helm chart to provide detail resource type of information.  Overall resource requirements or abstract requirements has limitation, such as overall resource requirements cannot provide enough detail pictures of the container, pod, cluster requirements.  Thus, most likely the deployment time would fail in many different scenarios.

    1. If so, is a full interpreter required?
    2. How much parameterization needs to get resolved?
    [Note] it all depend on the level of NFVO functionality needed.  It is assumed that NFVO would requires full interpretation and can ignore other part of helm chart where NFVO does not needed.

Note:

  1. Ericsson: SMO should not need to read Helm.  Some info may need to be duplicated in the ASD.   Risk of getting out of sync with Helm.  May need some compromise.   Needs more investigation.
  2. Nokia: Firstly: desire no duplication; May need some resource info in ASD but at a different level than in Helm but may cause complication.
  3. Information needed:
    1. Resources needed (how detailed??)?
    2. Connection points?
    3. Pod Topology?
    4. uService access points?
  4. What does the SMO do?
    1. Full NFVO+VNFM functionality?
    2. Simple LCM (Deploy)
    3. Mid LCM (Instantiate, Configure, Monitor, )

  1. Where are the required resources (CPU, Memory, Storage, ...) for each Pod described?
    1. Namespace quotas?
    2. Required (initial deployment) or Limit (runtime/scaling) Quota
    [note] the source of these information are coming from the helm charts.

    4. Where/How are the container images described?
    [note] the source of these information are coming from the helm charts and can be changed by deployment parameters.
     
    5. Where/How are the uServices described?
    [note] Further study of uServices represented on connection points, service access point, and external connectivity toward upper level (NSD).
  • No labels