Workflow Scheduling on Power Constrained VMs

With energy consumption being an issue of growing concern in large-scale cloud data centers, providers may wish to impose restrictions on the power usage of the hosts. This raises the challenge of operating cloud resources under power limits which may vary over time. Motivated by such a constraint, this paper considers the problem of scheduling scientific workflows in an environment where the number of VMs available is limited by a time-varying power cap. A simple scheduling algorithm for such cases is proposed and experimentally evaluated.