作者: Tianyi Chen , Xin Wang , Georgios B. Giannakis
DOI: 10.1109/GLOBALSIP.2015.7418248
关键词: Renewable energy 、 Data center 、 Real-time computing 、 Stochastic optimization 、 Operations research 、 Workload 、 Energy storage 、 Distributed data store 、 Intermittent energy source 、 Computer science 、 Quality of service
摘要: Recent advances such as high penetration of renewables, development energy-efficient cooling facilities, and flexibility distributed storage units, all call for a system-wide energy workload management policy future sustainable data centers. Leveraging stochastic optimization tools, unified approach is proposed allowing centers to adaptively respond renewable availability, efficiency variation, shift, price fluctuation under long-term quality-of-service (QoS) requirements. It rigorously established that the algorithm yields feasible near-optimal strategy without knowing distributions random workload, renewable, electricity processes.