[rm-wg] Updated Minutes

[David Snelling]

Folks,

Here is an update with people's names corrected.

Attendees:

Sergio Andreozzi
Marco Canaparo
Michele Carpenè
Dave Snelling
Steven Newhouse
Michel Drescher
Paul Strong
Erol Bozak


Introductions:

Restart after the summer. Dave and Paul have been chasing  
implementation activities related to the RM.

Next Steps:

1) Bring up the document to speed.

2) Sync with the GLUE WG next week.

3) Experiences with the model from Paul

	- Proposal: Remove ManagedComponent, retaining the capabilities.
		- Retain Manager and Pattern for the time being.
		- The manager component can be a managed component.
		- In CIM there is no explicit manager, but represented as services
		- How are the managing capabilities exposed - Not there in this  
diagram,
		     but they exist in the model.
		- Policy I still under study. At eBay it is not a GridComponent.
		- Agreed.
		- Action: Paul to update the UML.

3) Discussion of Grid Component State:
	- Operational State is behavioral.
	- Admin status is "where" the object is in it's external  
environment, e.g. workflow.	
		- Admin status is controlled externally
		- Lifecycle state is part of the Admin status.
		- The lifecycle states are dependent on the management capabilities  
that can be applied to the component.
		- There may be multiple views of the state.
		- Replace the component's notion of state with a lock and an owner.
		- The manager then provides the "value" of the Admin Status
		- What needs to be done?
			- Update the diagrams to reflect the rearrangement of properties.
		- Action: Paul to modify diagram accordingly.

4) State and Configuration
	- Configuration attributes describe the behavior of a component.
	- These are typically externally set
	- State attributes describe the component. These are typically  
changed internally.
	- Action: Paul to diagrams

5) The need to subclass relationships.
	- There is a small weakness in UML here.
	- RDF/OWL handle this nicely.
	- E.g. RunsInJVM <: RunsOn <: InstantionDepends <: DependsOn
	- Association classes cannot be subclassed.
	- Question: Should we move over to RDF or push UML a bit?
	- Agree that subclasses of relationships are needed.
	- Action: Paul to email Grady.
	- Sergio: To ask internally for other perspectives.

6) Action on all to provide feedback to the new diagrams.

7) Action on all to review GLUE documents.

https://forge.ogf.org/sf/docman/do/viewDocument/projects.glue-wg/ 
docman.root.drafts/doc14639

Moreover, the status and open issues of GLUE/RM integration is collected
in this page:
https://forge.ogf.org/sf/wiki/do/viewPage/projects.glue-wg/wiki/RM2GLUE




-- 

Take care:

     Dr. David Snelling < David . Snelling . UK . Fujitsu . com >
     Fujitsu Laboratories of Europe Limited
     Hayes Park Central
     Hayes End Road
     Hayes, Middlesex  UB4 8FE
     Reg. No. 4153469

     +44-208-606-4649 (Office)
     +44-7768-807526  (Mobile)



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.ogf.org/pipermail/rm-wg/attachments/20070928/e34a3dee/attachment-0001.html