Feature Set Summary

SDP PSI integration, testing, and enhancements. 

Work towards milestones:

• LOW PSI integration and testing
• End of bridging ATAM & progress review including lessons from the PSI

Miro planning board: https://miro.com/app/board/o9J_lZM70xo=/?moveToWidget=3074457353154812686&cot=14

Description

This feature Set continues the development and testing of a minimal SDP Operational System, with a focus supporting end-to-end system integration tests being carried at the LOW (and TBD MID) PSIs facilities.

By the start of construction (~end of PI10) the outcome of this work will be a minimal set of SDP services and workflows, that can be deployed as part of a distributed SKA software system, and can receive and capture data from upstream signal processing (CBF, PSS, etc).

In addition to this core capability, this system will need to provide an advanced operator with a well-defined mechanism to observe system behaviour and sufficient tools to debug and perform corrective actions to address issues without having to perform a complete system restart.

In addition to regular review at normal SAFe cadences, upon reaching the end of bridging, this system should be ready and sufficiently well documented (in Solution Intent) to be subjected to a fairly comprehensive architectural review, following an ATAM process or similar. 

Anticipated users

• PSI Operator
• SKA Architect

Use cases

• A PSI operator can view a list of validated workflows
• A PSI operator can select a workflow and easily modify its specified parameters
• A PSI operator can start a workflow via the telescope manager interface
• A PSI operator can view the status of a workflow
• A PSI operator can diagnose issues by means of logs, metrics, and traces, including QA metrics which diagnose issues with data being received.
• A PSI operator has a well-defined set of processes and supporting tools to recover from workflow failures.
• A PSI operator can cancel an SDP workflow
• ...

Outcomes

NFRs

• Stability:
  • The SDP system should not need to be restarted upon to recover from a partial failure (eg where a workflow has failed).
  • If an SDP LMC control service fails, recovery should take less than TBD seconds

Leading Indicators

• All parts of the SDP system are successfully integrated and tested using automated pipelines upon each code commit.
• All parts of the SDP system, meet minimal SKA software quality standards.
• New versions of the SDP system can be released on demand or at a minimal on a regular cadence that is shorter than a PI.
• A PSI operator can successfully execute an SDP receive workflow without developer intervention