An Empirical Approach of Performance and Energy-Aware Scheduling [PEAS] Mechanism in the HPC-Cloud Model

Cloud Computing provides the advantage of flexibility, elasticity, scaling, and customization to the HPC community as it attracts users that cannot afford to use the dedicated HPC infrastructure. HPC infrastructure is proven costly, as it requires upfront investment despite the advantage of processing the complex task. Interconnection of HPC and cloud environment creates the complex infrastructure for parallel computation and further creates a major issue in managing the makespan and energy performance trade-off. This research presents the PEAS (Performance and Energy-aware scheduling)-mechanism; PEAS is designed for parallel computation with task scheduling and optimal resource allocation at data centers. At first, a system model is designed for the parallel computing process; later, a novel and efficient scheduling algorithm is designed for task scheduling, and at last energy-aware mathematical model is designed for optimal energy utilization. PEAS are evaluated considering the HPC aware scientific workflow like cyber shake and montage workflow considering the evaluation parameter as Make span, Energy consumption, and Power utilization. Moreover, PEAS is proven to be more efficient than any other existing model available to date. © 2022 Seventh Sense Research Group®.