I have been trying to understand what's the added benefit of having an interface instance id and not only using the pointsource. I know the combination of pointsource and instance id is used by the interface to retrieve the pipoints specific to that interface. But this could easily be done through the pointsource alone, like suffixing it with an "instance id"; PS_1 and PS_2.
The only case i can think of is if the interface was split to two servers, each reading data from the same datasource, i.e. pointsource, for load balancing. Or having a PointSource that reflects the plant and have two or more interfaces reading from different sources, i.e. one opc and one rdbms. And this makes searching for pipoints easier than having to search with wildcards PS_*.
Are there other cases where it is used or is it for this purpose only?