| In recent years, with the rapid development of mobile Internet and the Internet of things, Internet penetration and business continued to grow. In order to ensure the service quality, Super Cloud Data Center (SCDC) emerges in the era of Internet. As the number of SCDCs rapidly increases, the problem of energy saving has become one of the most important problem in the industry.Software-Defined Networking (SDN) allows the control plane to be separated from the device, so as to realize the centralized control of the network. Existing researches about load balancing in data center don’t fully consider saving energy. Most researches concentrate on energy saving strategies in data center server nodes, which can’t maximize the advantages of the network structure of SDN. In this paper, based on Software-Defined Networking (SDN), we combine dynamic server load forecasting with scheduling strategy to optimize the data center energy saving strategy.In this paper, we propose a novel algorithm called Green Load Balancing of Data Center (GLBODC), which is a special, dynamic and self-adaptive strategy for cloud data center. In the case when the load of data center is in the middle or lower levels, the energy is saved when the load of the data center is balancing. We also propose an adaptive energy-saving strategy based on network traffic prediction of Data Center, using PSO algorithm to optimize the neural network model for network traffic prediction. The energy saving problem is solved through the study of SDN deployment of technology. Through SDN centralized control, the energy saving problem in joint control scheduling server nodes and switch ports is solved. Through this paper, introducing SDN network structure in the data center and saving energy through the load prediction algorithm can achieve better performance by deploying more optimized energy saving strategy, improve data center network equipment utilization and the overall energy saving performance of data center. |
PDF Full Text Request