CLAMP : application merging in the ECOIN context mediation system using the context linking approach

Abstract

Integrating data from heterogeneous data sources means dealing with context differences. That is, differences in the assumptions that are made regarding format and interpretation of the data. The Context Interchange (COIN) group has developed a formalism to describe the context assumptions of data sources and data receivers. An ECOIN application ties various sources together by being mapping them to a common ontology, or domain model. ECOIN applications allow the context differences between data sources to be resolved through context mediation. Users may then desire to merge together multiple ECOIN applications so that context differences across a much wider range of sources can be resolved with minimal additional effort. Accordingly, the work of this thesis focuses on the problem of merging ECOIN applications. The approach to merging discussed herein is termed context-linking. Context-linking allows the merging of ECOIN applications with minimal effort having to be spent on merging the ontologies of the two applications. This is achieved by employing a virtual approach to ontology merging that gives the appearance of a merged ontology. This thesis describes the merging process, presents a detailed case study, demonstrates the benefits of merging and presents the design for CLAMP, a tool that facilitates ECOIN application merging.