Energy Aware Genetic Algorithm for Independent Task Scheduling in Heterogeneous Multi-Cloud Environment

Abstract

Cloud datacentres contain a vast number of processors. The rapid expansion of cloud computing is resulting in massive energy usage and carbon emissions which has reported a substantial increase day by day. Consequently, the cloud service providers are looking for eco-friendly solutions. The energy consumption can be evaluated with an energy model, which identifies that, server energy consumption scales linearly with resource (cloud) utilization. This research provides an alternate solution to task scheduling problem which designs an optimized task schedule to minimize the makespan and energy consumptions in cloud datacenters. The proposed method is based on the principle of Genetic Algorithm (GA). In the context of task-scheduling using GA, chromosomal representation is considered as a schedule of set of independent tasks mapped with available cloud or machine in the proposed methodology. A fitness function is taken to optimize the overall execution time or makespan. Energy consumption is evaluated based on minimum makespan value. The proposed technique also tested upon synthesized and benchmark dataset which outperforms the conventional cloud task scheduling algorithms like Min-Min, Max-Min, and suffrage heuristics in heterogeneous multi-cloud system.