Dynamic Voltage and Frequency Scaling-aware dynamicconsolidation of virtual machines for energy efficient clouddata centers

摘要

Computational demand in data centers is increasing because of the growing popularity of Cloudapplications. However, data centers are becoming unsustainable in terms of power consumptionand growing energy costs so Cloud providers have to face the major challenge of placing themon amore scalable curve. Also, Cloud services are provided under strict Service Level Agreementconditions, so trade-offs between energy and performance have to be taken into account. Techniquesas Dynamic Voltage and Frequency Scaling (DVFS) and consolidation are commonly usedto reduce the energy consumption in data centers, although they are applied independently andtheir effects on Quality of Service are not always considered. Thus, understanding the relationshipbetween power, DVFS, consolidation, and performance is crucial to enable energy-efficientmanagement at the data center level. In this work, we propose a DVFS policy that reduces powerconsumption while preventing performance degradation, and aDVFS-aware consolidation policythat optimizes consumption, considering the DVFS configuration that would be necessary whenmapping VirtualMachines tomaintainQuality of Service.We have performed an extensive evaluationon the CloudSim toolkit using real Cloud traces and an accurate powermodel based on datagathered from real servers.Our results demonstrate that includingDVFS awareness in workloadmanagement provides substantial energy savings of up to 41.62% for scenarios under dynamicworkload conditions. These outcomes outperforms previous approaches, that do not considerintegrated use of DVFS and consolidation strategies.