Effective resource management technique using reservation pool

Abstract

Systems where resources are shared among the processes, Deadlock often occurs due to resource contention. The existing techniques that are used to deal with it either perform stringent testing or no testing at all. In both the cases either the overhead is substantial or the technique is inefficient. This paper aims to present resource reservation technique that will reduce the overhead for deciding about the resource allocation for system consisting of shared resources. It presents a technique for resource reservation namely Worst-Case Execution Time Based Resource Reservation (ETRR). The motivational examples illustrates that the proposed technique are capable of performing resource allocation without checking the safety sequence as proposed by Banker's algorithm. The overhead of resource allocation for the proposed techniques merely O(m) as compared to Banker's algorithm of O(mn2). Further, they are able to predict a deadlock more efficiently, when some of the other existing techniques fail to do so. The simulation results indicate that the average turnaround time of the ETRR is approximately 13% better than the existing Banker's algorithm.