Network Flow Scheduling Method Based On Sleep Awakening In Cloud Computing Environment

As the infrastructure of cloud computing,data centers usually run many various types of services,and there are varying degrees of resource idleness under uninterrupted work,which directly or indirectly causes a certain amount of waste of resources and energy consumption.From the perspective of the data center network,the paper designs a network flow scheduling method based on sleep wakeup to improve resource utilization and reduce energy consumption based on the premise of ensuring the needs of cloud users.The method has the following innovations:1.Propose a general framework to solve the problem of data center network energy consumption optimization.The framework addresses the problem that existing energy optimization schemes may reduce user service quality during the optimization process,introduces service satisfaction to strengthen the connection between network bottom-level scheduling and upper-level user application status,and proposes network flow classification based on different types of network flow bearer services Imagine that a penalty mechanism is finally set to reduce the frequency of excessive optimization of the data center network by the energy optimization framework.2.Propose a network flow avoidance method based on service satisfaction.This method addresses the problem that network resources cannot meet user service needs in a short time due to the sudden increase in cloud user requests.According to the network flow classification results given by the data center network energy consumption optimization framework,a certain strategy is used to target scheduling and classification Network flow,improve the priority of supporting user service network flow,ensure user service quality,and improve the utilization rate of network resources.3.Propose a dynamic sleep method for nodes based on service satisfaction.This method is aimed at the problem that the network fluctuations caused by the dormancy of the network nodes interfere with the user service quality.Firstly,the load situation of each network node in the data center is detected,and the network node with the lowest utilization rate is regarded as the object to be dormant.If the network can continue to meet the user's service needs,the current dormant object will be hibernated,and the dormant node will be awakened until the user service cannot be satisfied,to reduce network energy consumption while trying to ensure the user's service quality.Based on the general framework,corresponding energy consumption optimization methods are proposed for the two states of insufficient network resources and idle network resources.The two methods can be used independently to achieve optimization in one state,or they can be combined in parallel to achieve Multiobjective optimization.Through experimental verification,the network flow scheduling method based on sleep wake-up in the cloud computing environment proposed in the thesis can effectively improve resource utilization and reduce data center network energy consumption.