Attaining Semantic Enterprise Interoperability Through Ontology Architectural Patterns

Abstract

Enterprise Interoperability Science Base (EISB) represents the wide range of interoperability techniques that allow the creation of a new enterprise application by utilizing technologies with varied data formats and different paradigms. Even if one is able to bridge across these formats and paradigms to interoperate a new application, one crucial consideration is the semantic interoperability to insure that similar data is reconciled that might be stored differently from a semantic perspective. In support of this requirement, usage of ontologies is gaining increasing attention as they capture shareable domain knowledge semantics. The design and deployment of an ontology for any system is very specific, created in isolation to suit the specific needs with limited reuse in the same domain. The broad proliferation of ontologies for different systems, which, while similar in content, are often semantically different, can significantly inhibit the information exchange across enterprise systems. This situation is attributed, in part, to a lack of a software-engineering-based approach for ontologies; an ontology is often designed and built using domain data, while software design involves abstract modeling concepts that promote abstraction, reusability, interoperability, etc. The intent in this chapter is to define ontologies by leveraging software design pattern concepts to more effectively design ontologies. To support this, the chapter proposes Ontology Architectural Patterns (OAPs), which are higher-level abstract reusable templates with well-defined structures and semantics to conceptualize modular ontology models at the domain model level. OAP borrows from software design patterns inheriting their key characteristics for supporting enterprise semantic ontology interoperability.