2021-05-27 Meeting notes

Mahesh Jethanandani

• Mahesh Jethanandani brought up testing of O1 and VES interface. For O1, Bhanu Chandra K suggested that the simulator would have to be spun up to do the testing. If the simulator was up, then a curl command could be issued to validate the O1 interface. Something similar has to happen for the VES interface, where the shell scripts developed by OEM could be used to send events and validate that the VES interface is functional.
• Kiran brought up two questions.
  • What event types, in addition to the five event types that are currently supported, need to supported for the E-release?
    • Mahesh Jethanandani suggested that there is ongoing discussion on support of PM and FM event types which could be targeted for E-release. That discussion could add to the list of event types supported.
  • What event types need to be posted to the Kafka bus?
    • John Keeney (Ericsson EST) suggested that in ONAP all event types are posted to the Kafka bus, and it is up to the subscriber of the event type to then filter on the events.
    • John-Paul Lane suggested that there be multiple buses where different events are posted to different buses (??)

• John Keeney (Ericsson EST)suggested that "Day 0" configurations can be vendor specific configurations. That there may not be a standardized way to bring up all network functions
  • Mahesh Jethanandani, pnfRegistration is a known way for NF to let SMO know about their presence in the network. What will it take for the NF to send that to the SMO, so that the SMO can turn around and setup a NETCONF session with the NF?
• Alex Stancusaid that O-RU sets up a Call Home session with the SMO, so that the SMO can turn around and setup a NETCONF session with the O-RU. But that is not the same as pnfRegistration.
  • Note from Mahesh Jethanandani, Call Home was developed for a scenario where the NETCONF client is not able to setup a NETCONF session with the server because the server might be behind a firewall. But that may not be the case here. Would it be easier to just send a pnfRegistration event in that case?