基于SDN的集散式网络性能测量模型研究

发布时间：2018-01-12 08:37

本文关键词：基于SDN的集散式网络性能测量模型研究　出处：《河北大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着网络负载急剧增加,给网络的结构带来了很大的压力,软件定义网络(softwaredefined networking,简称SDN)顺势产生。准确的测量网络的性能是保证网络高性能运行的前提。在上述背景下,提出了一种基于SDN的拓展模型—集散式(centralized-distributed,简称CD)网络性能测量模型。对集散式模型的结构进行设计,实现了集中式的控制模式和分布式测量的过程。升级结构中的测量实体,对组件之间通信协议中的关键字进行制定,实现了模型中组件之间的通信和各结构组件的有限状态机中状态的转换。并且将CD模型应用到网络流量异常检测中,对全网络数据进行分布式在线收集,并且采用随机抽样方式对数据进行提取。针对提取到的数据,利用熵值进行异常判断。最后,针对提出的CD模型进行了仿真实验。实验验证该模型改善了网络流量的承载能力,提高了对数据的处理效率。可以准确的检测出网络中存在的异常行为。
[Abstract]:With the rapid increase of network load, it brings great pressure to the network structure. Software defines the network software defined networking. The accurate measurement of the network performance is the prerequisite to ensure the high performance of the network. Under the above background. In this paper, an extension model based on SDN, centralized-distributed, is proposed. The structure of the distributed model is designed to realize the centralized control mode and the process of distributed measurement. The key words in the communication protocol between components are formulated. The communication between the components in the model and the state conversion in the finite state machine of each structural component are realized, and the CD model is applied to the network traffic anomaly detection to collect the data of the whole network on line. And the random sampling method is used to extract the data. For the extracted data, the entropy value is used to judge the anomaly. Finally. Simulation experiments on the proposed CD model show that the proposed model can improve the carrying capacity of network traffic, improve the efficiency of data processing, and can accurately detect the abnormal behavior in the network.
【学位授予单位】：河北大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.06

【相似文献】