Research And Implementation Of Network Information Analysis System Based On Multi-machine

Along with the extensive application of the Internet, Internet is filled with alot of spam as well as a variety of network attacks. The key of creating a healthyand safe network environment is to filter the Internet information. Multi-patternmatching algorithm is usually used to filter spam and detect attacks. But mostmulti-pattern matching algorithms spend a lot of memory to minimize time. Thecommon server may not be able to provide filter system with enough memory. Tosolve the single server that can not work when dealing with large-scale patterns,the information filtering system based on multi-machine is proposed in this paper.Firstly, this paper explains the principle of AC algorithm and WM algorithm.With the same prefix relationship between pattern string, AC algorithm jumpsfarther distance when mismatching. WM algorithm introduces the ideas of badcharacter and HASH technology to improve the efficiency of the match. Inaddition, this paper presents interfaces of AC and WM algorithm for themulti-machine system, such as data transmission interface, query interface.Secondly, this paper presents patterns partitioning algorithm for themulti-machine architecture. As for the design of multi-machine, this paper showscommunication protocol and interface for calling AC and WM algorithm. Patternspartitioning algorithm is the important part of the filtration system. Patternspartitioning algorithm affects the whole system memory allocation and theresponse time of system.Finally, this paper shows the system function and performance test. Forfunctional testing, tests include partitioning patterns, system query function, so asto verify the validity of the system function. For performance test, tests includequery time under this distributed system and system in the single server.