"It appears functions such as video retrieval and exporting are done by the HP (Autonomy) "VMS system", while distribution of and access of the CCTV video streams are done in the Vidsys PSIM side."
That sounds troublesome but not surprising. Unless Autonomy is OEMing a real VMS, you could be foregoing quite a number of valuable VMS features. Vidsys is clearly not a VMS and neither is Automony. Plus, the VMS market is quite mature, and the types of VMSes used in projects like yours tend to be quite advanced.
I'd ask them: Are you (Autonomy) OEMing a true VMS like Genetec, Milestone, etc.? If so, which one? If not, how can they be competitive with those offering?
As for your question:
"How/where a common view of entities such cameras, camera access, camera hierarchy is held and maintained?"
That would be in the PSIM, in their approach. Problem is usually PSIMs have very rudimentary video controls / playback because (1) they are not real VMSes and (2) they are constrained to only deliver the most basic controls that every VMS / recorder they integrate with supports.