Update Propagation Technique for Data Grid

Data replication is a well known technique used to reduce accesses latency, improve availability, and performance in a distributed computing environment. An asynchronous replication is a commonly agreed solution for the consistency of replicas problem. Update propagation using a classical propagation schema called the radial method suffers from high overhead of the master replica while line method suffers from high delay time. This paper presents a new asynchronous replication protocol called Update Propagation Grid (UPG) which especially for a wide area distributed Data Grid. Updates reach other replicas using a propagation technique based on nodes organized into a logical structure network that enables the technique to scale well for thousands of replicas. Restructuring operation is provided to build and reconfigure the UPG dynamically. An analytical model is developed; communication cost, average load balance, and average delay time have been analyzed. The technique achieves load balancing and minimizes the delay for file replication in Data Grid.