Performance And Cost Optimization For Data Center Driven By Distributed Generations And Energy Storage Devices

With the coming of big data era,data center develops faster and faster.The scale and performance of data center grow quickly.Behind the convenience data center brings to us,the problems of high energy consumption and high pressure on environment it causes are becoming more and more serious.Recently,fuel cell and gas turbine gather lots of attention as green energy.Making use of fuel cell and gas turbine can help data center cope with the high energy consumption problem and environment problem.However,adjusting the power output of fuel cell and gas turbine needs response time which may incur the mismatch of power supply and power demand.Energy storage device is the key enabler as the backup power,which not only improves the energy efficiency but also shaves the peak power of data center.This paper studies on a green data center driven by distributed generation and lead acid lithium battery.To cope with the mismatch of power supply and power demand,we propose power capping and power shaving technology in the power supply perspective and make use of DFS and task deferring technology in the power demand perspective to balance power demand and power supply.Moreover,short task priority scheduling strategy helps to increase the performance.Shaving the peak power of distributed generation and battery can reduce the cost.Finally,we can save 37% cost in the expense of 5% performance reduction in average.