Design and Architecture of the FDBS Prototype INFINITY

This paper focuses on the architecture of our prototype supporting the integration of heterogeneous data. The main characteristics of our system are its extended schema architecture and the generic translation approach based on a mapping language. At first, we introduce the schema architecture as well as the essential properties of our mapping language. One of our major contributions is the execution model which describes the system dynamics when data is derived from different sources and combined/converted to the specified application view. As the second essential contribution, the paper presents an optimization mechanism, called context management, to address the deficiencies resulting from an object-oriented interface on top of relational database systems.