Performance analysis of greedy Load balancing algorithms in Heterogeneous Distributed Computing System

摘要

The Load balancing problem on Heterogeneous Distributed Computing System (HDCS) deals with allocation of tasks to compute nodes, so that computing nodes are evenly loaded. Dynamic load balancing problem to assign tasks on HDCS is presented as a linear programming problem to minimize the makespan. Different greedy resource allocation algorithms are presented for load balancing on HDCS with tasks as Expected Time to Compute(ETC) matrix. The relative performance of the four heuristic algorithms under different circumstances has been simulated on two different HDCS using in house simulator. The simulation results show that the greedy based scheduling policy has the significant impact on the system heterogeneity.