Energy-efficient scheme based on user task characteristic in virtual cloud platform

Focus on node cloud computing environment due to the randomness of the user task arrival caused a large number of idle energy consumption. To solve this problem, a dynamic energy-efficient scheme in virtual machines based on user task characteristic is proposed. First, according to the amount of user tasks arrival and the expectations of task completion time, user tasks characteristic information table is established. Thus it can predict the request arrival probability and user expect to reach the finish time. Then, according to user expectations of completion time, psychological expectations QOS value of the user is obtained, and the threshold based on the integration of energy-saving value is dynamically adjusted. Finally, it can achieve energy-saving by dynamic integrating of the user task arrival rate and closing idle node. Experimental results show that the method can reach a better balance of performance and energy saving to meet user requirement.