Ip Network Traffic Capture And Classification Techniques

With the widespread of various network applications ,the internet have expanded gradually with geometric isomerism while it is became more complex . At the same time, people's requirements on their network performance in daily lives became more and more large, lot of distributed applications Required excellent network capabilities and service quality when they use the network.Network must be stable, reliable and optimize the performance to provide services that make applications more efficiency. At the same time, the application users need to verify whether the services available to meet their performance requirements. In recent years with the expansion of the network size, a lot of energy being used for the understanding and definition of network behavior, a large number of complex applications and network characteristics cause the analysis of the network became more and more difficult. In order to achieve this target, we not only need to design high-performance network protocols and develop high-performance network equipment, but also need to achieve a reasonable distribution of network load and ensure critical business performance of network through traffic engineering .the Traffic Engineering achieve reasonable network load distribution and to ensure that critical business performance is also crucial. Therefore, it is extremely important for the development of IP networks to set up ahigh-speed IP network base on the high-speed IP network traffic monitoring activities we carried out.This dissertation attempted to discussed the existing network traffic measurement techniques and analysis methods,and designed a high-performance packet capture platform which improve the efficiency of packet capture by studied the traditional Packet Capture platform and theory of Zero Packet Copy System.Meanwhile,this dissertation also improved the Recursive Flow Classification (RFC) algorithm which require larger memory consumption after analyzing the classical multi-domain Packet Classification Algorithm. Finally,we designed a HTTP protocol message packet analysis module to analyzed the specific HTTP protocol packet belong to the TCP/IP protocol.