摘要
针对云计算、大数据等互联网应用规模不断扩大,新应用的发展对传统网络提出更高效集中的网络管理需求、高效灵活的组网需求。文章对在软件定义网络架构下的数据中心网络路由方面展开研究,提出基于流分类的软件定义数据中心网络路由算法,使用改进的路由算法,在Fat-Tree架构上建立流量模型并与传统的ECMP算法进行性能分析和对比。结果表明,文章中提出的路由算法能够在提高链路利用率的基础上提高网络吞吐量和降低分组端到端时延。
With Internet application scale such as cloud computing, big data, expanding unceasingly, the development of new applications for traditional network makes a more efficient centralized network management requirements, highly efficient and flexible network requirements. This paper spread a study in the software-defined data center network routing aspect, putting forward software-defined data center network routing algorithm which based on traffic classification, proposing an improved routing algorithm, building the traffic model on the Fat-Tree architecture and comparing with the traditional ECMP algorithm. The results showed that the proposed routing algorithm can improve network throughput and reduce packet end-to-end delay on the basis of improving link utilization.
作者
赖香武
彭大芹
黄德玲
刘艳林
Lai Xiangwu Peng Daqin Huang Deling Liu Yanlin(Communication and Information Engineering College of Chongqing University of Posts and Telecommunications, Chongqing 400065, Chin)
出处
《无线互联科技》
2016年第24期38-40,共3页
Wireless Internet Technology