An Energy-efficient Scheduling Strategy on Processors Frequency Determination

The blind pursuit of low processor execution frequency does not necessarily get an effective energy saving effect in Dynamic Voltage Scaling(DVS) energy saving implementation of real-time system.The energy consumption of real-time task is not only related to the power,but also related to its execution time,to reduce voltage/frequency results in the increasing of execution time thus induces the increasing of energy consumption.In this paper,through static analysis of energy consumption model based on DVS in data center of real-time system,a new approach to determine the optimal initial execution frequency for processor to complete transactions is proposed.Instance data calculation results show that based on the optimal initial frequency to complete the transaction can save about 30% of energy consumption than just using the maximum processor frequency.