To Shift Tasks or to Shift Energy by ESDs? an Economical Scheduling for Cloud Data Center.

In this paper, we propose a novel scheduling scheme to minimize the electricity cost for cloud data center by making the best of time-varying electricity prices of the power grid. In our model, the tasks may have different arrival time, execution time and deadlines, so that the energy cost can be reduced through task shifting within their deadlines. In addition, the energy cost can also be reduced through energy shifting when its price is low using energy storage devices (ESDs). However, it is challenging to decide whether to shift tasks or to shift energy using ESDs considering the fluctuating prices of the electricity. Note that, there are energy losses during charging, discharging and self-discharging of the ESDs, which can not be ignored. We solve the problem in two steps: 1) schedule the tasks based on the electricity prices within a certain time; 2) model the scheduling of energy into a constraint optimization problem and solve it using CPLEX. Experiment results show that our scheduling can significantly reduce electricity cost for a cloud data center.