The Research Of Performance Tuning On Teh Prefix Email Filtering System

As the development of Internet, email has become an important tool of daily contacting. Email provides convenient communication. However, many applications related to email cause big troubles in the Internet. The most of all, the flooding spam.As the product of email filtering technology, the prefix email filtering system performed well at the entry time in the market. Yet, the filtering system is not good enough in both the decoding module and filtering module. To solve the core problem of performance improvement, this thesis includes the following parts:Firstly, facing the flooding spam, this thesis summarizes all features of the spam and email filtering technologies.Secondly, with the netfilter/iptables running on Unix, this thesis provides the method of network security improvement for filtering system.Thirdly, this thesis not only gives the way of filtering English words and simple Chinese character(GB2312), but also gives a set of methods to filter traditional Chinese characters(GB5 or GBK). Then, the thesis extended the filter ability on Email type without much cost.Fourthly, based on the tuning of identifying algorithm for the encoding method of mail body, the differentiation of the whole mail body and accessory parts to the Email, and the Email head data procession, the thesis improves the ability of Email decoding module, enhances the ratio of finding the spam in all the Emails.Lastly, this thesis referred to the vicinity category classification and introduced this technique in prefix Email filtering system. After this, the Email filtering system can filter pornographic counteraction and spam related to advertisement. Till this time, this article has some test results which showed the high call ration to pornographic spam, and there is not much obvious reduce in the efficiency of filtering system.