Based On The Distributed Nodes Of The Internet Data Stream Analysis System

By the developing of the network, the communication of information becomes more and more convenient. But at the same time, there are very great negative effects also, such as the spread of various kinds of violence, superstition, reactionary, pornography and some other illegal information, and the leaking of secret information. In view of these phenomena, the analysis of network data-stream is very important. But the increased use of new network applications, all kinds of communication protocols become more and more complex, and use variety algorithms of data-compression and data-encryption. To meet the current and future development trends, the traditional filtration technology which based on TCP/IP and simple keyword filter can not solve the problems effectively.In this thesis, a distributed network data-stream analysis system has been researched and analyzed. This system based on traditional data capture technology analyzes the data-stream of common application layer network protocol by the link features and data format of every protocol.This system's main feature is:1).Distributed system architecture. The distributed computers working as analysis node, make the system simultaneously on a number of network data-stream analysis.2).Dual-mode data access model. Both from the hardware and software from the network data packet capture, so the system can be adapted to different network environment.3).A high performance, scalable data-stream analysis kernel. It becomes more flexible and easy of protocols by different network protocols to be unified in the form of Dynamic Link Library Calling.