基于邮件路径地理属性分析的垃圾邮件过滤算法

Spam filtering algorithm based on geographic E-mail path analysis

提出一种基于邮件路径地理属性分析的邮件过滤算法GEPA(geographic E-mail path analysis)。首先提取邮件命令报文包含的路由信息,并以此为基础构建邮件路径子集;其次采用一种高效的地理属性映射方法进行地理信息映射;接着对路径中节点的地理逻辑关系背离情况进行分析用于过滤垃圾邮件;最后从中国大陆某骨干网边界路由器的一条链路上(该链路跨越地理边界)采集邮件流量以验证算法性能。研究表明,GEPA识别的垃圾邮件约占邮件总量的13.9%,且算法在执行速度和内存开销等方面具有较好的性能,能够满足实时邮件过滤的需求。

A geographic E-mail path based algorithm called GEPA （geographic E-mail path analysis） was proposed to allow network administrators to cut off spam traffic on E-mail delivery. The algorithm first extracted route information to build E-mail path subset, and then uesed an effective method mapping IP addresses or domain names of nodes in an E-mail path into geographic information. Further, the algorithm detected spare by their geographic information deviation, using E-mail traffics from a link of backbone border router in China, which crosses the country boundary of China, the performance of GEPA algorithm is evaluated. The experimental results indicated that a 13.9% reduction of E-mail can be achieved with method. The results also showed GEPA was effective and practical which can be implemented in a massive traffic environment handling over millions of mails every day with small memory consumption.