I've just begun populating our Oracle Enterprise Repository (220.127.116.11.0), and I have a query about values used for Asset Type and Asset Function ...
When I submit a Service asset, it appears in the editor under Submitted, Pending Review
Once Accepted, it moves to Submitted, Under Review
and once reviewed, updated and registered it moves to the "Registered" folder
BUT, the Registered folder contains a range of sub-folders which correspond to the values of AssetFunction
I'd prefer to see sub-folders based on AssetType - i.e. Services, Applications, Interfaces etc.
Does anybody have any comments ?
I also have a query about whether to govern at Service level or at Service Operation level - for finer-grained impact analysis
Are there any documents that discuss Best Practice or recommendations ?
Regarding the granularity of governing - don't you have some more coarse-grained services that eventually invoke more fine-grained services ? That's what we have. Ideally you'd want to trace the dependencies between them all I'd think.
Start small, get your buy-in from your business people and get your roles and responsilibilties and amendments to current processes sorted out as well.. The repository can help with governance but also there is the People and Process side of it that unless is tackled and you get buy-in from your stakeholders your repository won't get used and you will end up going back to spreadsheets and word-of-mouth.
Thanks again for your response. I do indeed have the situation you describe ... coarse-grained services calling finer-grained services.
I have recorded both types of service in OER as Service assets (using different values of Asset Function to differentiate between the two types of Service), and I've created References/Is Referenced By relationships between the Services.
That works well for Services with a single operation, but if the target Service is a Data Service with a range of CRUD-type operations I can currently only record the relationship between the calling service and the target Data Service.
There is no indication of which operation(s) the calling Service uses.
This is what I was referring to as fine-grained impact analysis.
I could change the meta-model to add an Operation asset and model the relationships that way, but are there other options that you're aware of ?
Hi, no I'm not aware of any way to do this. You could see what the harvester does if you automatically harvest it. I'd be interested to see what relationships it comes up with. I suspect it might just be the more coarse-grained relationship rather than the explicit operation.