期刊名称：International Journal of Grid and Distributed Computing
印刷版ISSN：2005-4262
出版年度：2016
卷号：9
期号：1
页码：25-36
DOI：10.14257/ijgdc.2016.9.1.03
出版社：SERSC
摘要：In order to improve the performance of data transmission in SDN, this paper proposes a load balance solution scheme by taking advantage of the global network view of SDN. We collect 4 load features from each transmission path. These features are bandwidth utilization ratio, packet loss rate, transmission latency and transmission hops. By using this 4 load features, BP Artificial Neural Network model is trained to predict the integrated load for different path and to choose one with least load as the data-flow transmission path. The contrast experiment results show that load balancing strategy proposed in this paper can select more rational transmission path for data-flow and achieve 19.3% network latency decrease at most.
关键词：Software Defined Network; Load Balance; BP Artificial Neural Network
