DADTA: A novel adaptive strategy for energy and performance efficient virtual machine consolidation

Abstract

Large-scale virtualized data centers are increasingly becoming the norm in our data-intensive society. One pressing challenge is to reduce energy consumption in such data centers for edge computing deployment, which would have flow-on effects on reducing the operating costs and carbon dioxide emissions. Dynamic virtual machine consolidation is an effective way to improve resource utilization and energy efficiency. In this paper, a comprehensive strategy is proposed, which is based on time-series forecasting approach. In this strategy, specific adjustment of threshold is applied to adapt the dynamic workload. We then use a real-world dataset (i.e. workload trace in Google) for evaluation, whose findings demonstrate that our strategy outperforms other benchmarks.