基于序列特征的网络流分类方法研究

Research of network flow classification based on sequence characteristics

如今,对网络流量中各种应用进行准确分类和识别已经变得越来越重要,针对目前流量分析研究的不足,本文综合国内外相关研究成果,提出了在双向动态网络流模型的基础上,采用细粒度的Packet-Level序列特征属性对流量进行分析,建立序列特征属性与网络流类型之间的关联关系,实现了一种高效的、与端口无关的网络流分类方法。

Accurate classification and identification of network traffic in a variety of applications has become increasingly important. Considering the shortcoming of traffic analysis research, based on two-way dynamic network flow model, the paper researches on the relationship between the sequence characteristics of network flow and network applications by analysing packet-level attributes to achieve an efficient network traffic classification.