Binary Information Press
With the increasing power consumption in the storage system and the effect of global greenhouse, power efficiency has attracted a wide spread interest. The system components with different states have different performance and power cost. To address the issue of power efficiency, deciding when to execute component state transitions and perform which transition is critical. A Stochastic Power Management (SPM) scheme based on random Iterative-workload Parameters Estimation (IPE) hence is proposed. Upon the iterative estimation which can promptly and dynamically follow-up the changing parameters of workload, the model of Markov power management and the optimization equation of power consumption meeting the user demands are constructed and then the power management policy is achieved by solving the optimization equation with LP algorithm.