The Study On The Energy Saving Strategy Of Reserved Virtual Machine In The Cloud Environment And Batch Arrival Queuing Model

With the rapid development of cloud computing,more and more people choose to complete their work in the cloud data centers.The huge energy consumption generated by the cloud data centers is non-negligible.In order to reduce energy consumption and realize green cloud computing,we propose energy saving strategies for reserved virtual machines in the cloud data centers,establish batch arrival queueing models,evaluate and optimize the system performance.At first,for the cloud data centers that need to better ensure the quality of customer service,we propose an energy saving strategy of reserved virtual machines based on synchronous standby.When the number of cloud tasks in the system is relatively small,the reserved virtual machines are in standby state to reduce energy consumption.Once the number of cloud tasks in the system exceeds the threshold,the reserved virtual machines will be active state immediately providing services to ensure the user response performance.Secondly,for the cloud data centers which need to better ensure the energy saving effect,we propose an energy saving strategy of reserved virtual machines based on synchronous hibernation.When the number of cloud tasks in the system is relatively small,the reserved virtual machines are in a periodic state of hibernation to save more energy.When the number of cloud tasks in the system exceeds the threshold,the reserved virtual machines will enter the active state after the ending hibernation to alleviate the load of the cloud data center.Then,considering the multi-task user request in the cloud environment,introduce the batch arrival mechanism and establish the multi-server queueing system.In view of the energy saving strategy of the reserved virtual machines based on synchronous standby,we establish the queueing model of batch arrival and partial servers synchronous close.In view of the energy saving strategy of the reserved virtual machines based on synchronous hibernation,we establish the queuing model of batch arrival and partial servers synchronous multiple vacation.The steady-state distribution of the system model is derived by using Gauss-Seidel method.Through numerical and simulation experiments,the system performance of energy saving strategy is evaluated and the effectiveness of energy saving strategy is verified.Finally,considering the relationship between the blocking probability of a task,the average response time and the system energy saving rate,respectively,we establish the system cost function for two kinds of energy saving strategies to reveal the influence of the system threshold,the sleep timer parameter and the service rate on the system cost.In order to improve the accuracy of searching,we improve the teaching-learning-based optimization algorithms,the optimization scheme of the system parameters is given under different user request arrival rate.