Ontology-Mediated Data Migration: Deriving Migration Rules by Reasoning on Schema Descriptions

Migration of data across information systems is a knowledge intensive task: the definition of mappings between systems requires knowledge of the source and target (relational) schemas and their interpretation of the shared domain. Moreover, direct schema mappings need often to be re-defined for each new migration instance, in order to accommodate the variations caused by the change of systems and representation conventions. A possible solution to such problems is the use of an intermediate ontological model, that can be used as a lingua franca for the description of schemas, by defining mappings from and to the ontology. While this helps in making explicit the semantics of the schemas, the problem remains on how to extract a direct mapping from source to target schema from this intermediate representation. In this paper, we present our ongoing work in building an ontology-based migration system in the scenario of banking information systems. In the architecture of the system, an ontology defines an intermediate semantic description for the source and target schemas. We introduce a reasoning method for the automatic extraction of migration rules starting from the semantic descriptions of the schemas. The procedure for computation of migration rules is then implemented via reasoning over an Answer Set Programming encoding.