Study of Network Traffic Recognition Based on Unsupervised Learning Method

Modern network management requires more and more precise to network traffic of a great variety of applications. However, the traditional network management software can only obtain monitoring of the three and four layers network performance and checking of network traffic. The achieved information is simple and sparse. Therefore, to meet requirement of modern network, it should separately monitor each network traffic application. There are mainly two methods: one is deeply packet inspection and the other is based on features of data flow. But the two methods have their own advantages and disadvantages. To implement real time monitoring of traffic of application layer and finding ability to new application at the same time, a new method is proposed to combine two above methods with unsupervised learning to identify current sophisticated traffic data generated from various emerging network-based applications. Finally the conclusion and the research work in the future are presented.