Energy-Aware Resource Prediction in Virtualized Data Centers: A Machine Learning Approach

The availability and high utilization of the vast computational power in the current technological era results in a high electrical power consumption rate. This rapid growth and demand for computational power lead to the creation of large-scale data centers which have high power utilization requirements thus resulting in high operational costs. Based on these observations and analysis of machine learning for virtualized cloud data centers' this paper proposes a machine learning approach for energy-aware resource prediction in a virtualized data center environment. The proposed method utilizes the polynomial regression model to predict the likely power consumption and the number of machines that are physically needed based on the daily workload. The proposed model is also discussed.