Performance and Power Management for Cloud Infrastructures
Saving power without interfering both performance and workload is a crucial task. Different mechanisms, algorithms and strategies are presented, compared and discussed. From the live migration and consolidation of virtual machines to the creation of management frameworks which encapsulate the whole workflow. Afterwards, virtual power states which allow virtualization-management-policies to save power are presented. Every strategy is discussed while specific issues are highlighted. The comparison shows that there are opportunities for significant power saving while meeting customer Service-Level-Agreements (SLA) to ensure the Quality-of-Service (QoS).