Smart Power Management in Data Centers Using Machine-Learning Techniques

Data centers contribute more than 2% of the carbon emissions across the globe. The data centers comprising various servers and computing systems with diverse performance profiles, present a challenge to job schedulers and maximum usage of the virtual machines that are running on them to maximize the infrastructure’s benefits. The prime objective of this chapter is to consider energy consumption and data-center performance in order to evaluate, analyze, and present them, since there is an increase in the consumption of energy in data centers making use of machine-learning regression-based algorithms for prediction and then carrying out virtual migration (VMs) techniques to make maximum utilization. In this chapter, the work summarizes the development of significant power savings, a powerful energy analytic and monitoring of real-time power consumption in data centers. The design also aims in measuring and quantifying the carbon dioxide (CO₂) emissions precisely. This supports the idea that deliberate access and strategies could be successful in addressing energy and climate challenges, hence lowering enervation.