Designand Implementation Of Network Traffic Monitoring System Based On Multicore Processor And DPI

With the advent of the Internet age, more and more people are involved in the complicated network activity, their personal privacy and property information may be exposed to the network. In order to guarantee the security of the network environment, we need to monitor and manage the information on the network accurately. Large network traffic monitoring needs higher computational performance and more complex control strategies, traditional network traffic monitoring solution hardly meet performance or functional requirements. In response to the above situation, this paper proposes a network traffic monitoring solution based on MIPS and deep packet inspection technique.Firstly, this article introduces the OCTEON multi-core network processor. The content covers the processor architecture, task procedure, fundamentals and usages of functional hardware units and multi-core programming strategies and so on. Then this paper present the common techniques and application scenarios of deep packet inspection, and put forward the use of multi-core processor on deep packet inspection.Then this paper summarizes several common traffic monitoring techniques and methods, analyzes several common network application packets, and proposes a packet application identification method based on keyword probability model. On the basis of packet application identification, a packet application prediction strategy is presented based on5-tuple and temporal locality.On the basis of the above work, this paper gives a network application framework based on OCTEON multi-core processor, which includes functional grouping, pipeline design and stages division. Then a network traffic monitoring system is designed and implemented on the hardware platform and application framework. Finally, packet content inspection, application identification and prediction, monitoring functions are tested. In addition, packet capturing and storage capabilities also have been tested respectively. These test results demonstrate that the monitoring system can effectively inspect the keywords carried by the packet on the network, and be able to identify the application of network traffics, and also be able to process the identified packet by pre-set operations. The system runs stably, correctly in the practical network environment.