Design And Implementation Of Network Traffic Identification System Based On DPDK

Nowadays server network bandwidth and hardware performance continue to improve,and Internet services have rapidly changed from traditional HTTP,SMTP,FTP to P2 P,VoIP,streaming media and other services.The new types of services represented by Thunder,Baidu Cloud,Tencent Video,etc.occupy a large amount of bandwidth resources while being used,and cause network congestion.In addition,some companies have a very urgent need for employees to manage their online behavior at work.Faced with these problems,the identification and control of network traffic and applications will help to ensure the security and transparency of communication.It is very valuable for netowrk traffic control.In view of the above problems,this paper first studied the high-speed packet capture technology that can adapt to the high-speed network environment;on this basis,the high-speed packet capture mechanism was designed and implemented;with the dynamic ports and hidden ports being widly used,The identification technologies based on port,Deep Packet Inspection(DPI)or Deep Flow Inspection(DFI)were studied and analyzed.This paper compared the application scenarios,advantages and disadvantages of the three methods.Finally,The paper selected the DPI recognition technology,designed and implemented a fast and effective DPI traffic recognition engine.Then,combined the packet capture mechanism based on DPDK and the traffic recognition engine based on DPI.a high performance,low-coupling network traffic system is designed.The requirements analysis and overall design of the network traffic analysis system are completed.The paper expounded the implementation and related methods of each functional module of the system.Finally,the system is tested for function and performance.It verifies that in high-speed network environment,the system has the ability to identify network traffic in real time,which reflects the high recognition rate and high performance of the system.