When we need to integrate two systems that share a common process we align the semantic models (smodels) of each system.

Since all systems in the process map to the same semantic dictionary (sdictionary) we can draw equivalences between the low-level syntax of each system. JUMP does this automatically by aligning the higher-layer semantic objects of the sdictionary. The corresponding DB or XSD schemas that were mapped into the semantic objects in the smodel are considered equivalent.

The smodels allow information elements in each schema to be correlated so that the discrete information fields in the different schemas that have the same or similar meaning can be aligned. The Crosswalk builds the table or "maps" that show these relationships.

Any two systems in the workflow that need to exchange data create similar Crosswalks for metadata alignment.

Each system on the bus views the smodel from its own perspective data model. It requires no direct information from any remote service regarding their low-level schema. All a service requires of another service is their smodel to dynamically generate a transformation.