基于网络资源的kNN网络流量分类模型的研究

A Study on Network Traffic Classification Model of kNN Based on Network Resources

为了快速有效的识别网络流量中的攻击行为,根据同一类应用对网络资源的消耗具有相似性,不同类应用对网络资源的消耗具有差异性,提出了一种基于网络资源的kNN流量分类模型。该模型可以从多种网络资源中选取三种属性作为特征,采用改进的kNN算法进行流量检测,对网络中的各种流量特别是攻击类流量尽可能快速并有效的识别,给网络防御的启动提供重要依据。实验结果表明:改进的kNN算法在保证识别率的同时有效提高分类的速度。

With the substantial growth of network users and network applications,the attack behavior which is hidden in the traffic seriously affects the quality of network service.How to identify the network traffic quickly and effectively is one of the keys in network security.According to similarities in the consumption of the network resources by the same sort of applications,and the differences in the consumption of the network resources by different applications,this paper proposes a network traffic classification model of kNN based on network resources.The model can be used to select three attributes as its own feature from a variety of network resources,and it uses improved kNN algorithm to detect traffic.It can also identify all kinds of traffic in network as fast and effectively as possible,especially the attack traffic.It can be used as an important basis for the enablement of network defense.The experimental results show that the improved kNN algorithm can improve the speed of the classification effectively while ensuring the recognition rate.