Flexible Framework for Data Integration and Update Propagation: System Aspect

Dejima architecture is a framework for sharing data and controlling update over multiple databases. Each peer autonomously manages its own database and collaborates with other peers through views. In this paper, we describe the design concept and the mechanism of update propagation in the Dejima architecture. We explain how to identify the affected peers for a given update and propagate the update to the affected peers in a single transaction. We also claim that Dejima architecture is a flexible framework that combines both aspects of two former approaches for data integration, the global schema based one and the peer based one.