Energy-Aware Virtual Machine Consolidation for Sustainable Cloud Data Centers

Abstract

Cloud data centers consume vast amounts of energy, and much of it is wasted because servers run at low utilisation yet still draw substantial power. Consolidating virtual machines onto fewer physical hosts, so that idle hosts can be powered down, is a leading strategy for reducing this waste, but it must be balanced against the risk of overloading hosts and violating service-level agreements. This paper proposes an energy-aware virtual machine consolidation approach that migrates virtual machines to minimise the number of active hosts while respecting utilisation limits that protectperformance. The approach was evaluated by simulation against a no-consolidation baseline and a conventional threshold-based policy across arange of workloads. The proposed approach reduced energy consumption byroughly thirty percent relative to no consolidation and outperformed thethreshold-based policy, while keeping service-level violations low andrequiring fewer migrations. The results show that energy-aware consolidation can substantially improve the sustainability of cloud data centers without an unacceptable cost in performance, making it a practical tool for greener cloud operation. KEYWORDS: cloud computing, virtual machine consolidation, energyefficiency, data centers, green computing, live migration