Burke, S (Stephen) said:
Thinking a bit more, I think the diagram is right and the table is wrong, e.g. consider an MPI job which might be running in more than one ExecutionEnvironment.
... and while I'm looking for problems with these relations, the Manager table says Resource.ID <<abstract>> 1..* A manager manages zero or more resources. where the multiplicity should be 0..* (you may just want to publish a Manager to advertise what software is installed). Similarly for Resource we have: Manager.ID <<abstract>> 1 A resource is managed by a manager. This is maybe related to the idea that Resource is supposed to be a child of Manager. However, I don't really see why it shouldn't be allowed to publish Share and Resource but not Manager. Fundamentally the Resource (hardware) and Manager (software) descriptions are peers, neither is dependent on the other so I think it should be possible to publish either one without the other. Stephen -- Scanned by iCritical.