I am working on an aggregating server, which connects to multiple OPC UA servers, of which some have identical information models and some slighly differing models. The aggregating server only uses standard types and its address space is modeled to a very specific hierarchy, different from any of the underlying servers that are to be aggregated. The aggregating server must search for objects of a specified type from the underlying servers, and map some of the nodes belonging to each found object to its own information model and hierarchy. Not all child nodes of the found object are aggregated, and the nodes that are are reformatted to a completely different hierarchy. Some properties may need to be added and some attributes may need to be changed. This should all be done automatically.

So basically I need a way to define a set of rules which tells each aggregating client how to handle each object that needs to be mapped (e.g. what object type in the aggregating server corresponds to the found object type, which variables/properties should be mapped or ignored, what should be the structure of the newly created aggregated object and where in the address space it should be placed etc.). I was wondering if anyone has done something similar before, or if there is a known technology already designed for this (or that could be utilized for this)? I can’t seem to find information on this specific subject.