Decentralized data management framework for Data Grids

A new class of data intensive applications has led to increased demand for cost-efficient resource sharing approaches. Yet, providing efficient access to widely distributed data for large numbers of users poses considerable challenges. Most existing Grid systems are centrally managed, thus hindering their scalable expansion. We introduce a new distributed, adaptive, and scalable middleware that provides transparent access to data in Data Grids. Our approach relies on dynamic techniques that adapt replica creation to continuously changing network connectivity and users' behavior. Results from simulations and deployment of our middleware show that our solution provides better data access performance than static approaches. (C) 2006 Elsevier B.V. All rights reserved.