Reflective data sharing in managing Internet databases

Abstract

The popularity of the Internet has generated an explosion in the number of accessible information sources. In addition, recent advances in wide-area networking have led to a push for a logically-unified, yet physically distributed, information repository accessible through the Internet. The architecture of data resources and sources should be scalable and able to accommodate hundreds and thousands of databases. These are called Internet databases. This has to be achieved against a backdrop of accommodating database autonomy and bridging their heterogeneity with reasonable overhead. In addition, manual and evolution management needs a careful design as ad hoc management is clearly intractable in a highly dynamic environment as the Internet. Therefore, there is a need for supporting a self-adjustable model of evolution. We propose an architecture that has the ability to reflect upon its own state and to decide when to evolve. The use of flexible constructs like service links, coalitions, and co-databases provide a dynamic and portable model for eliciting data sharing. Further, we propose an implementation model using CORBA and Web related technologies.