Efficient task scheduling in cloud networks using ANN for green computing
Hadi Zavieh,Amir Javadpour,Arun Kumar Sangaiah
DOI: https://doi.org/10.1002/dac.5689
2024-01-20
International Journal of Communication Systems
Abstract:In this study, we propose ANNDB, a novel approach utilizing artificial neural networks (ANNs) for energy‐efficient task scheduling in cloud environments. Through extensive simulations, ANNDB outperforms existing methods (WPEG, IRMBBC, and BEMEC) with improvements of 13.81%, 8.62%, and 9.74% in energy efficiency and 3.93%, 4.84%, and 4.19% in power efficiency, respectively. Our findings highlight ANNDB's potential to optimize cloud computing environments, reduce energy consumption, and promote sustainable computing practices, making it an attractive option for businesses seeking greener and more efficient task allocation solutions. Summary Recently, there has been a growing emphasis on reducingenergy consumption in cloud networks and achieving green computing practices toaddress environmental concerns and optimize resource utilization. In thiscontext, efficient task scheduling minimizes energy usage and enhances overallsystem performance. To tackle the challenge ofenergy‐efficient task allocation, we propose a novel approach that harnessesthe power of Artificial Neural Networks (ANN). Our Artificial neural network Dynamic Balancing (ANNDB) method is designed toachieve green computing in cloud environments. ANNDB leverages the feed‐forwardnetwork architecture and a multi‐layer perceptron, effectively allocatingrequests to higher‐power and higher‐quality virtual machines, resulting inoptimized energy utilization. Through extensive simulations, wedemonstrate the superiority of ANNDB over existing methods, including WPEG,IRMBBC, and BEMEC, in terms of energy and power efficiency. Specifically, ourproposed ANNDB method exhibits substantial improvements of 13.81%, 8.62%, and9.74% in the Energy criterion compared to WPEG, IRMBBC, and BEMEC,respectively. Additionally, in the Power criterion, the method achievesperformance enhancements of 3.93%, 4.84%, and 4.19% over the mentioned methods.The findings from this research hold significant promise for organizations seekingto optimize their cloud computing environments while reducing energyconsumption and promoting sustainable computing practices. By adopting theANNDB approach for efficient task scheduling, businesses and institutions cancontribute to green computing efforts, reduce operational costs, and make moreenvironmentally friendly choices without compromising task allocationperformance.
telecommunications,engineering, electrical & electronic