Improvement of Anti-Spam Performance on Postfix

• Main Authors: Nien-Wang Chang, 張年旺
• Other Authors: Wei-Chih Hsu
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/04793445402738883175

碩士 === 國立高雄第一科技大學 === 電腦與通訊工程所 === 94 === Abstract Among the spam related research journals, most of the researches are focused on comparison of databases and filtering analysis. The related researches are range from spam filter performance analysis and mail terms destrbute database efficiency research. However, only small divisions of researches are focused on its related network agreement. This thesis is paying attention to the current filtering system plus two unique functions: Automatically White List (AWL) and Manual Black List (MWL) functions to improve the filtering results. The thesis can be separated into two parts. The first part is analysis of current technological advantage and disadvantage plus efficiency comparison .The second part, we provide a better filtering mail technique which can increase efficiency. As we understand, currently there is no system can 100% filter out all the spam and still need update database from the vendors. However, the databases update can never faster than produce the spam. Also, each vendor provides database which may not fit the company’s filtering rules. Therefore, error filtering out mails occurs frequently. The only way to improve filtering efficiency is to use quick and accurate filtering rules. Due to previous reasons, this thesis suggests to increase the filtering Spam mail technique on Postfix of the mail system. This technique is basic on Bayesisn algorithm and Gerylisting program update database concept. From the testing lab, we check the testing data for orginal and proposed. In according to filter add AWL and MWL or not. The results get decreasing network loading and better anti-sapm performance than mail system present.