A Replica Placement Algorithm in Mobile Grid Environments

Abstract

With the advent of high bandwidth third-generation mobile networks and other wireless networks, grid computing has migrated from traditional parallel and distributed computing to pervasive and utility computing based on the wireless networks and mobile devices, which results in the emergence of a new computing paradigm named Mobile Grid. Data replication is widely used in mobile grid environments to reduce access cost and improve data availability. Therefore, replica placement becomes one of important challenges for the replica management in mobile grid environments. This paper addresses the QoS-aware replica placement problem in a tree based mobile grid environments. At the same time this paper proposes a two-step solution for this problem: the first step is a bottom-up dynamic programming approach for placing replicas to satisfy the QoS requirements of mobile users and workload constraint of replicas; then, a binary search based algorithm is used to finish the k replica placement problem in mobile grid environments. Experimental studies show this solution is feasible and efficient for the replica placement problem in mobile grid environments.