Modeling Demand Response Capability by Internet Data Centers Processing Batch Computing Jobs

Modeling Demand Response Capability by Internet Data Centers Processing Batch Computing Jobs Electricity cost has become a big concern of commercial cloud service providers with the rapid expansion of network-based cloud computing. Locationally, dispersed large-scale Internet data centers (IDCs) that underpin the cloud have increasing impact on the regional electricity market with their skyrocketing energy consumption. In this paper, the electricity market participation and accordingly the demand response capability of an IDC is defined as its temporally and spatially shiftable electricity demand quantities for processing delay-tolerant central processing unit-intense batch computing jobs. The demand response capability of the IDC is obtained by the proposed electric demand management solution. Price-sensitive and cooling efficiency-enabled batch computing workload dispatch with the objective of minimizing electricity cost is realized by dynamic IDC server consolidation and scheduling with virtual machine live migration technology. Numerical simulations show the effectiveness of the proposed demand management solution in IDC energy consumption and electricity cost reduction.