The Automation meetings alternate times each meeting to accomodate the global team.

Agenda

* Reoccurring agenda items:

• Recap of previous meeting AutomationMeetings20111208
• Are Deployment/Dev Ops scenarios considered complete?

• Main page cleanup: AutomationHome
• Draft Spec discussion: AutoSpecificationV1
  • Overview of where things are currently at
  • Resource relationship diagram
  • Environment/Parameter attribute representation
  • Result Contribution attribute representation
  • Do the proposed V1 artifacts satisfy the V1 scenarios?
• Additional workgroup members interested in participating in spec authoring
• Next meetings:
  • 12 January

Minutes

Attending: Michael Fiedler, Pramod Chandoria, Rich Rakich, Pete Steinfeld, Max Vohlken, Louis Paul, Paul McMahan? , Dave Brauneis, Scarlett Li, Charles Rankin, Dan Berg, John Arwe, Sheehan Anderson, Srimanth Gunturi

• Review of Dev Ops and Deployment discussion from AutomationMeetings20111208
  • Request to not consider the scenario closed until colleagues still on holiday return
  • Identified a need for an Appendix to provide guidance on items not explicitly addressed by the specification, but which require consideration when implementing service providers and clients
    • Environment tear-down considerations - differing requirements drive needs to tear down quickly or retain
    • Retention of automation results - depending on scenarios, some may require long-term persistence for audit or problem re-create
    • In composite or orchestrated automations, the dependencies between results and requirements for persistence could require planning
• Review of AutomationHome
  • No major comments. Confirmed intention is to address all scenarios in a single specification.
• Start of review of AutoSpecificationV1
  • Covered resource/relationship diagram
    • Clarification on the concept of template Automation Plans (cf. Deployment and Dev Ops scenarios).
    • Agreement that template Automation Plans are Automation Plans which exist to be cloned and customized, but never run.
      • Can still be exposed by OSLC, but never expected to have requests or results.
      • Can contain partial specification of input parameters which are further specified/customized when cloned and when Automation Requests created.
    • Discussion on relationship between Results, Requests and Plans
      • Re-visited the topic of Request persistence. Confirmed some providers will have very transient Request artifacts
      • Confirmed the correct ReportsOn? (name could change) relationship is Results -> Plans. "Users" want to see all/filtered results for a plan.
      • Result -> Plan direction of the relationship allows for scalability - no Plans with 000s of links to results. Use OSLC query to find Results of interest.
  • Started to review attributes
    • Main topic was the parameter attribute
      • Discussion on how V1 spec will allow for "any" remote or inlined parameter structure - format not specified in V1. Requires consumer to understand parameter structure to some extent.
      • Possible need for separate input and output parameters. What are the roles of parameters as we move from Plan->Request->Result.
      • Is there a need to add parameters as things progress? For use by downstream automations? Consensus seems to be yes - this can clarify what parameters are to be used/were used in each stage.
      • TODO: Continue the discussion. MichaelFiedler will start a discussion on the mailing list

• Next meeting: 12 January at 1PM Eastern US time.