The improved greedy gang scheduling by minimizing context switch condition

Abstract

An effective resource management, called scheduling, is essential for the performance of large-scale distributed systems. One scheduling technique is gang scheduling, performing scheduling for parallel jobs in gang type. In this paper, a new algorithm for gang scheduling is proposed. This method aims to reduce the average response time of gangs by increasing the serviceability of gangs in the shortest execution time possible. The performance of the proposed algorithm is examined and compared to the basic gang scheduling algorithm within the simulation. The results of the simulation indicated that the response time of the proposed modification compared to the basic method is reduced up to 40% with low values of multiprogramming and high pressure of workload (short inter-arrival time) in Adapted First Come First Served and Largest Gang First Served policies.