(Luis) Issue #3 for draft-ma-opsawg-schedule-yang – Grouping for sequence of events

[luismcontreras]

In some cases it is useful to create scheduled sequences of actions. For instance:

• Section 3.2 of draft-contreras-opsawg-scheduling-oam-tests proposes OAM test sequences. The sequence can be used for creating the individual set of unitary OAM tests, and then for running such sequence along the time. The latter can be covered by current approach in draft-ma-opsawg-schedule-yang, but not the former.

• In draft-contreras-tvr-alto-exposure the assume use case is the one of a NOC populating scheduled topological changes for instance due to the upgrade of a node (HW, or SW) or any generic network maintenance intervention. It can be assumed that after the works in the network the initial situation will be reverted. Thus, having a sequence of actions (i.e., schedule change, and reversion of the change) can be more straightforward than creating two different scheduled actions.

So, having a grouping for sequence of events can be beneficial.
Such a grouping can be something like:
grouping sequence:
+-- time-zone-identifier? sys:timezone-name
+-- sequence? Identityref
+-- number-of-actions? Identityref
+-- action
| +-- date-time-start? union
| +-- time-zone-identifier? sys:timezone-name

(The grouping description is probably not complete)
It can be considered that an action starts just after the previous action finishes.

[boucadair]

Are there generic "sequences" that we can define in a common model?

+-- sequence? Identityref

Also, how this can be reused by external modules where the actual set of actions will be defined?

[luismcontreras]

I think defining generic sequences of actions could be useful, yes. Maybe defining in a generic way "sequence" and the number of actions in the sequence, while external modules detail each of the actions. Would it make sense?