Several modifications

[nigel-r-davis]

I notice I am not a formal reviewer, which is fine. But wanted to leave a few comments.

I recognize that this release is not going to be enforced in the RIA and that it is a stepping stone to the more formal release later in the year (which will be enforced in the RIA). On that basis my comments are not blocking as we have time to refine the model etc. I took a look at the changes via the Yang diffs and read the Yang (I am apparently converted!).

I looked at the following areas:

• tapi-alarms:

  • Many of the new fields do not have description. We need to deal with this as we approach the main release later this year
  • I still have a major concern with "alarm-raised-time" etc. The original notification based alarm and the streaming alarm do not have these fields. I still see no justification for their inclusion. I have raised this concern on a number of occasions.
  • occurrence-times is not explained. It cannot (meaningfully) be simply the number of times the detector has gone active (from day zero). It seems to be a fragment of a problem model. We do need to discuss this in greater detail.
  • The alarm has a very traditional flavor (severity etc.). I hope we can progress beyond this for the main release.
  • The alarm structure does not merge the streaming alarm and the notification alarm. I think that would have been a better first step
  • suspect interval from TCA highlights one of a range of potential issues, it would be worth us assessing this in general

• tapi-oam:

  • I was expecting to see a reference from oam-job to cep (for the NCM case). I assume that this is still to be done.

• tapi-streaming:

  • The module is not up to date... I will create a pull request to cover this before the RC1 is released.