Research And Application Based On The Fpga Network Traffic Behavior Analyzer

With the rapid development of the Internet business, the new protocols andapplications continue to emerge, people have become increasingly demanding of highquality of network service. Accurate and rapid analysis of network traffic behaviorand effective traffic management is the basis of a high quality network. Networktraffic behavior analysis first requires accurate identification of the traffic class, buttraditional identification methods in the identification of new protocols and newapplications increasingly ineffective.After the failure of the traditional traffic identification methods, manyresearchers try to use data mining techniques for traffic identification. Previousstudies have compared traditional port-based and payload-based analysis recognitionmethod with clustering-based identification method; the result shows that the newmoth has advantages of scalability, identification accuracy and performance. Thedisadvantage of such methods is not mature enough, the choice of clusteringalgorithm and characteristics impact on the identification accuracy greatly. To solvethis problem, based on clustering algorithm we research how to select the appropriatefeature and algorithm parameters to improve the identification accuracy, and make theidentification model can support online real-time identification to meet the needs oftraffic management.There is an online operating process in real-time when do the traffic analysis.Traditional platform as the PC has the shortcomings of low efficiency and poorreal-time. It is difficult to be developed for a new application because the flexibility ofASIC platform is very low. Even the NP platform, it only accelerate normal networkoperate while the business algorithm is still run with non-real-time software mode. Inrecent years, the functionality and performance of FPGA is rapidly developed. FPGAis widely used in various fields. Compared with traditional platform, FPGA is good atflexibility, high efficiency, and speed. Therefore, we research how design and realizeof a FPGA-based network traffic processing hardware platforms and how implementtraffic identification algorithm on this platform.Finally, this paper researchs how to implement a traffic filtering system on the FPGA platform. The traffic filtering system is connected to the network link,real-time analysis of network traffic behavior to identify traffic class and prevent theunnecessary flow, release normal flow. To demonstrate the efficiency and speedadvantages of FPGA and to meet the real-time demands for traffic identification, westudy how to design and implement network traffic filtering function with IP core. IPcore runs as a hardware gate circuit, it has high performance.