An analytical model for electricity-price-aware resource allocation in virtualized data centers

Costs associated with the power consumption and cooling requirements of servers in a data center are significant. For this reason, a lot of research efforts in the area of power management have been devoted toward greening data centers and clouds in the last years. This work starts from the observation that the main target of a DC manager is not only to save energy for the environment, but also reduce operational costs. To this purpose, the paper focuses on a single server management, and proposes a joint power management and resource allocation strategy that modulates the number of active VMs on a given server according to both the current workload and the current value of the virtual machine power cost, in order to minimize the total server management cost, while matching the SLA with the customers. An analytical model is proposed to support design of the resource allocation controller parameters.