[Back]

B. Schreder:
"Legacy Datasource Interoperability for the Semantic Web";
Supervisor: S. Dustdar, F. Rosenberg; Institut für Informationssysteme, AB Verteilte Systeme, 2013; final examination: 2013-10-07.

This master thesis discusses the problem of integrating legacy data sources with other connected applications and data sources, one of the key issues of processes such as Enterprise Application
Integration. This work examines one possible solution to this problem by leveraging semantic technologies, specifically, by mapping between ontologies which describe the legacy data sources. This solution necessitates the mediation and alignment between a data source´s or
application´s data format and the format prescribed by the central, shared ontology. Utilizing ontologies is assumed to provide a flexible and light-weight solution to the data integration problem.
In this thesis several semantic mediation approaches are surveyed, including the requisite preprocessing steps such as schema lifting and normalization. Using the Mapping Framework of the University of Karlsruhe as the underlying architecture, a comprehensive mediation tool has been developed, which supports the creation and execution of mappings between different ontologies. Several new methods of importing legacy data sources, such as relational databases and XML data, are developed and introduced as new services for the mediation tool. Finally, a number of case studies and application scenarios for these mediation services are detailed and their applicability and flexibility are evaluated.