Continuation of TMC integration with Dish Software

Single source of truth: YAML spec, derived from ICD, becomes the "ICD".
"ICD" needs to be agreed between Dish guys and SKA SAFe programme.

Agile integration: Johan's utility allows you to assess a device agains a set of criteria which does not include the total ICD content of dsh-lmc-prototype device content. Concept of "minimal YAML file" that Anton mentioned. We assume today's ICD will change. Indicate what is agreed and what is not.

Subset determined by what behaviour and characteristics the MVP needs in a certain PI.
Gerhard agrees with this.

Then assess simulator against this. Gerhard calls DishMaster python device an emulator. Emulator needs to honour YAML file and exhibit the behaviour required of it by the MVP to perform its functions. TANGO Simlib suggested as option to implement.

Maybe have another YAML, full agreement, in addition to YAML that says how the simultor should look.

Step 1: Create YAML file and get agreement on it in SKA SAFe project, and share with INAF for feedback. This is only a sub-set of the full interface. This YAML file tells us what the Dish Simulator should have on its interface: selection criteria is what the MVP needs for PI7 (repeat in future) - Gerhard's sequence diagrams can help define this. Also should support MeerKAT CAM for MeerKAT - SKA Mid Dish integration, based on TM MID to Dish ICD.

Step 2: Create a Dish simulator that honours the YAML file. Test Dish simulator against YAML file.

Change TMC if required.

Integration: Aggregation that at least includes TMC and Dish simulator/emulator. TMC tests should pass (tests in tmc repo that are integration tests).

Focus on step 1 now, and then move on to step 2, replan.

Previos description:

During PI6 SP-235 enabler caused dsh-lmc-prototype to be containerised and deployed as part of the MVP, limited integration tests between TMC and dsh-lmc-prototype and assessment of the prototype against the SKA Control System Guidelines. During that it was found that there are discrepancies between the TMC and the dsh-lmc-prototype.

Discussion 18 May 2020 with g.leroux

• scope work around what interaction the MVP needs from TMC and dsh-lmc-prototype.
• Focus more on exposing integration problems so as to inform fixing problems.
• do "standalone" component testing of TMC and of the dsh-lmc=prototype against the TM to Dish ICD to ensure that we have some objective evidence to manage discrepancies:
  • interface testing against ICD: maybe contract-based testing (g.brajnik, can this be done as part of SP-629?). Karoo team wants to explore a "validation tester" for interface testing asgainst ICD and some aspects of CS Guidelines that may also have more general use.
  • Behaviour testing of TMC and of dsh-lmc-prototype as BDD tests.  Part of this is to extend dsh-lmc-prototype sub-system test coverage from trivial mode change test to include other commands and attributes.
• Deployment: try deploying Dish leaf nodes standalone for integration flexibility.

Integration tests involving TMC and Dish LMC to ensure Dish LMC behaves as required to support the MVP. This will inform decision to replace existing Dish simulator with the INAF Dish LMC.

• Fix/resolve the TMC-DSHLMC some issues identified by ----SAR-99--

• Replace the DishMaster simulators in skampi with the real DSH LMC DishMaster (maybe uncommitted objective).

• Compile a list of changes required to dsh-lmc-prototype to correct misalignment to ICD, CS Guidelines, TMC.