Traffic Prediction And Energy-saving Scheme In Data Center Network Based On SDN

In recent years,as technology improves day by day,a number of data centers come to the fore all over the world.They provide IaaS(Infrastructure as a Service),PaaS(Platform as a Service),SaaS(Software as a Service)three services to customers.Nowadays,many commercial enterprises,research institutes and governments look at influence of these data center from a lot of parts such as security,available and environment-friendly.Because a data center consumes huge energy,emits a large amount of carbon dioxide and very expensive,a lot of institutes mentioned above focus on improving utilization ratio of data center by researching computing system,software design,power supply and cooling system.Based on this,we discuss usability of neural network traffic prediction model for the traffic in data center.Then,we propose a dynamic energy-saving scheme and a framework based on OpenFlow in data center,simulation results testify that the scheme not only balances traffic capacity and energy saving better than typical energy-saving strategy,but also decreases response time and extra energy while OpenFlow switches changing status.