Hi all, I've asked similar question before with ILM 2007, but this topic is still not clear for me, and I think it should be discussed in some beginners examples. What I want, is to manage reference attributes between different object types. The main problem is that these objects are mostly from different MA and hence it was nearly impossible to manage reference attributes between them as they couldn't point to themselves. And here comes the new version FIM 2010 which contains the FIMService DB, where all of these objects are stored. Here is some example to describe the "problem": I have one MA that is projecting person objects. This object should contain a multivalue reference DN attribute e.g. myCertificates, that should point to another object type in MV (certificate) containing some info about certificates that belongs to that person. Another MA is projecting these certificate objects into metaverse, where this object should contain another reference attribute "owner" that should point to the person object who owns this certificate. My question is how can I fill these attributes in the most simple way. Of course there is some connection between these objects (person's ID attribute contains string that should correspond to another attribute of certificate object and vice versa). I still don't understand why there is no possibility to fill these values in some MV extension or why there is no option to create FIM Service MA extension project. I think that all of this can be accomplished by combination of MPR and some action WF but I would need some examples of workflows with filling the reference attributes. Hope it sounds clear. Thanks for any advice or a clue.

I'm afraid, there is no "simple way" to do this and you will have to add more details to your scenario description.To address you scenario, I wrote an article (Design Concepts: About Reference Attributes) that summzarizes the basic requirements for managing reference attributes.In essence, there is no way to manage reference attributes if the referencing and the referenced objects are in different connector spaces.This is due to referential integrity.If you can't get both parts of a reference relationship into a connector space, you will have to dereference them.I hope, this helps to shed some light on this.Cheers,Markus Markus Vilcinskas, Knowledge Engineer, Microsoft Corporation

Thanks for shedding some light on this. But anyway, my point was, as you said, that I need to have referencing and the referenced objects in the same CS, so why not to use CS from FIM Service MA. Unfortunately, I cannot involve some extension code on this type of MA, which is a little bit shame. But today I realized that on the FIM portal I can actually define a group with computed members. If I would be able to create these groups dynamically (by provisioning of objects from MV) with some formula defining computed members, it would be really great and at least simplify most of the referencing tasks. If this is not possible right now, I think the potential of added FIM Service is not fullfilled. I hope that you see my point. Cheers, Tom

The objective of the FIM CS is to function as mirror of the metaverse.This is why there is no need for rules extensions.Calculating data relationships has never (including ILM) been in the scope of the synchronization engine.These are application level operations that should be implemented in a connected data source.The FIM data store is an example for such a data store.So, instead of modeling your data relationships in extension code, you could, for example, write a web service client that does this for you.Regarding reference attributes, the scope of the synchronization engine is "just" to enforce referentiell integrity across connected systems.Cheers,Markus Markus Vilcinskas, Knowledge Engineer, Microsoft Corporation