Uncertainty Management for Multiple Data Centers in Transactive Electricity Market: A Cloud Federation Approach

The efficient operation of data centers necessitates a resource management strategy to cope with stochastic incoming workloads. Meanwhile, with an increasing amount of fluctuated data center loads integrated into the distribution network, it is critical to alleviate the deviation between actual and forecast data center demand to ensure economic grid operation. To this end, this paper proposes a cloud federation framework dedicated to managing workload uncertainties of multiple data centers as a coalition while mitigating their power deviations. In this framework, a model that can obtain the desired workload adjustment towards zero deviation for each data center is proposed. Upon receiving the workload adjustment willingness of all data centers, a cloud federation broker is responsible for determining the optimal workload migration among data centers with complementary deviation patterns. The simulation results validate the effectiveness of the proposed cloud federation framework, as the total penalty cost decreases by 35.2% and the average power deviation decreases by 57.5% compared to no federation case.