Research On Energy-Saving Virtual Machine Scheduling Strategy In Cloud Environment

For the past few years,more and more services and applications had run in the cloud,cloud service providers have had to expand the size of cloud data centers.Large-scale data centers not only consume huge amounts of electricity,while the release of large amounts of carbon dioxide will also have a certain impact on the global climate.Therefore,in recent years,data center energy-saving technology has been widespread concerned and researched.To ensure that the cloud service provider's own interests and cloud user's satisfaction,users and cloud service providers need to sign the service level agreements(SLA,Service-level Agreement).This paper mainly studies how to achieve the purpose of energy saving in the data center through reasonable scheduling of virtual machine under the premise of ensuring that SLA is not violated.In order to study how to reduce the energy consumption in data center.In this paper we analyze the system model in virtual machine scheduling,the mathematical model of virtual machine selection and placement problem in virtual machine scheduling,the energy consumption model of single physical node in cloud data center,the virtual machine migration costs and the cost of SLA violation.We conclude that it is possible to reduce the total costs by setting a reasonable physical node's CPU threshold.In the fourth part of this paper,we analyze the advantages and disadvantages between static threshold and dynamic adaptive threshold detection on an overloaded physical machine and their application scenario;we analyze the shortcomings of using the greedy algorithm to select migration virtual machines and propose an algorithm which based on dynamic planning to select migration virtual machines;We analyze the virtual machine placement algorithm based on Best Fit Deceasing,and propose an algorithm to solve the virtual machine placement by using the artificial fish swarm algorithm.Experiments show that the selection and placement algorithm of migration virtual machines proposed in this paper have good energy saving effect and can guarantee the SLA violation rate is within the acceptable range.