Over load detection and admission control policy in DRTDBS

Abstract

Today's Real Time System (RTS) is characterized by managing large volume of distributed data making real time distributed data processing a reality [1, 2]. The demand for real time data services is increasing in many large scale distributed real time applications. The transaction work load in DRTDBS may not be balanced and the transaction access pattern may be time varying and skewed. Hence, computation workload on large scale distributed system can lead to large number of transaction deadline misses [3, 4]. Hence, efficient database management algorithm and protocol for accessing and maintaining data are required to satisfy timing constraints of transaction supported applications. In this paper, an algorithm has been proposed for admission control in DRTDBS consisting of local controller and global load balancer working at each site, which decide whether to admit or reject the newly arrived transaction. The simulation results show that the new algorithm successfully balances the workload in DRTDBS [5, 6]. © 2011 Springer-Verlag.