Distributed Heterogeneous Relational Data Warehouse In A Grid Environment

Abstract

This paper examines how a "Distributed Heterogeneous Relational Data Warehouse" can be integrated in a Grid environment that will provide physicists with efficient access to large and small object collections drawn from databases at multiple sites. This paper investigates the requirements of Grid-enabling such a warehouse, and explores how these requirements may be met by extensions to existing Grid middleware. We present initial results obtained with a working prototype warehouse of this kind using both SQLServer and Oracle9i, where a Grid-enabled web-services interface makes it easier for web-applications to access the distributed contents of the databases securely. Based on the success of the prototype, we proposes a framework for using heterogeneous relational data warehouse through the web-service interface and create a single "Virtual Database System" for users. The ability to transparently access data in this way, as shown in prototype, is likely to be a very powerful facility for HENP and other grid users wishing to collate and analyze information distributed over Grid.