A Dynamic Ontology Mapping Architecture for a Grid Database System

Abstract — Most large-scale heterogeneous distributed computing systems, such as Grids, rely on Service Oriented Architectures (SOA) to interact with others in different platforms and computing languages. However, we still need to solve the semantic heterogeneity problem of data; we must interpret the data from different systems in some semantically related ways. Ontologies are the most common and well-accepted methodology to handle this problem at multiple levels of granularities across different systems. Nevertheless, using ontologies in a dynamic environment, such as a Grid, to share some common concepts is still a challenge. It is difficult to keep a static mapping between ontologies; the corresponding semantic mapping changes must occur consistently. Therefore, we adopt the concept of Tuple Space and propose a flexible approach for managing ontologies in a Grid. It enables systems and users to interoperate semantically and dynamically by sharing and managing the concepts and semantic ontology mappings in a flexible approach. I