Energy Aware Workload Scheduling Metrics for Execution of Parallel Application in Heterogeneous Cloud Computing Platform

The modern workload application is generally being executed on Heterogeneous (CPU-GPU) Cloud Computing (HCC) environment. Minimizing execution cost with high quality of service is most important considering economic perspective of Cloud Service Provider (CSP). The energy consumption plays a major part in cost and also makespan, meeting application deadlines. This work addresses the challenge involved in scheduling parallel workload under heterogeneous cloud computing for minimizing energy and makespan meeting task deadline prerequisite. Existing model are efficient in minimizing either energy or makespan; thus, resulting higher service provisioning cost. In addressing the research problem, this work present Energy Optimized Scheduling (EOS) for parallel workload application in heterogeneous cloud computing platform. Experiment results shows the proposed model is efficient in minimizing energy and makespan in comparison with existing workload scheduling approach. The result shows an average energy reduction of 23.22%, overall power consumption reduction of 85.06%, and makespan reduction of 78.2% for montage workflow in comparison with energy minimized scheduling.