Enhanced Load Balancing Strategy with Migration Cost on Heterogeneous Distributed Systems

Heterogeneous Distributed Systems (HDS) provides service to users executed on a set of geographically heterogeneous distributed resources with dynamic nature. In HDS, load balancing approaches has a vital role in tasks allocation for avoiding resources overloaded or underloaded. Over the past decades, these approaches have been developed to improve objective parameters. The mapping problem of independent Batch of Tasks (BoT) on computational machines with heterogeneities to optimize various quality of service (QoS) in distributed system has been proven as NP-complete. In this paper, we propose an Enhanced Load Balancing Strategy with Migration cost (ELBSM) to optimize load imbalance on nodes in the system. We evaluate our work through comparative performance evaluation using randomly generated batch of tasks by using MATLAB 8.5.0. Experimental study suggests that the proposed work prevails over existing strategies in terms of considered parameters under study.