面向企业的反垃圾邮件网关系统的设计与实现

发布时间：2019-06-09 17:33

【摘要】：垃圾邮件占用了大量的有限的存储、计算和网络资源,影响和干扰了用户的正常工作生活,同时钓鱼邮件是特殊的垃圾邮件,它欺骗收件人账号、口令等信息,危害正在逐步增大。垃圾邮件充斥着互联网,大大的降低网络运行效率,要从垃圾邮件的源头和传播途径上治理垃圾邮件是非常艰巨的。基于垃圾邮件对企业用户的种种危害,为解决企业员工在使用企业邮箱中免受垃圾邮件、钓鱼邮件的困扰,研究反垃圾邮件相关关键技术的功能实现,设计并实现面向企业的反垃圾邮件网关系统是十分必要的。本文在分析课题相关关键技术的研究现状的基础上,重点研究邮件过滤技术及过滤分类器的更新机制。本文的主要工作如下:1、本文采用基于机器学习反馈的文本分类方法winnow算法,对接收的邮件根据分类器自动进行内容过滤判定,同时分析钓鱼链接的特征,通过基于URL的域名的钓鱼邮件过滤方法,并且结合URL的域名白名单相似度的方法,实现对邮件中钓鱼链接的综合过滤。2、本文提出了一种基于用户反馈和蜜罐技术的邮件分类器更新机制,利用蜜罐技术收集垃圾邮件,同时通过用户对反垃圾邮件网关所隔离的垃圾邮件和企业邮箱中的正常邮件的反馈情况,提取邮件集,为定期自动更新分类器提供邮件集合。通过使用最新的邮件样本集来训练分类器,使得分类器能更好的判定新邮件的类别。3、研究了面向企业的反垃圾邮件网关系统的前后台各模块设计方案,前台包含了系统管理、基础配置、隔离区管理、日志管理等功能模块,后台包含了主控模块、邮件接收、邮件原文解析、邮件内容分析、钓鱼邮件检测、邮件转发、蜜罐技术、分类器更新训练等模块,详细阐述了系统各模块的具体设计、实现及测试,并给出了典型的应用场景。综上所述,本文设计并实现了一种面向企业的反垃圾邮件网关系统,通过对课题的关键技术的研究,不仅从提高垃圾邮件过滤准确性的角度来提高邮件过滤网关的识别能力,而且设计了基于用户反馈和蜜罐技术的邮件分类器更新机制,通过不断的学习训练,从而保证分类器达到识别能力最优,实现网关系统总体的垃圾邮件识别率的提升。通过本文的设计研究能够对进出企业内部的邮件实时过滤,能够在一定程度上减少企业网络带宽占用,确保邮件服务器稳定运行,减少垃圾邮件、钓鱼邮件对企业网络的威胁。
[Abstract]:Spam takes up a lot of limited storage, computing and network resources, which affects and interferes with the normal working life of users. At the same time, phishing mail is a special spam, which deceives the recipient account, password and other information. The harm is gradually increasing. Spam is full of the Internet, which greatly reduces the efficiency of network operation. It is very difficult to control spam from the source and transmission of spam. Based on the harm of spam to enterprise users, in order to solve the problem that employees are free from spam and phishing mail in the use of enterprise mailbox, the function realization of key technologies related to anti-spam is studied. It is very necessary to design and implement an enterprise-oriented anti-spam gateway system. Based on the analysis of the research status of related key technologies, this paper focuses on the mail filtering technology and the updating mechanism of filtering classifiers. The main work of this paper is as follows: 1. In this paper, the text classification method winnow algorithm based on machine learning feedback is used to automatically filter and determine the received mail according to the classifiers, and the characteristics of phishing links are analyzed at the same time. Through the phishing mail filtering method based on URL domain name and the method of domain name whitelist similarity of URL, the comprehensive filtering of phishing links in mail is realized. 2, In this paper, an update mechanism of mail classifiers based on user feedback and honeypot technology is proposed, which uses honeypot technology to collect spam. At the same time, through the feedback of users to the spam isolated by the anti-spam gateway and the normal mail in the enterprise mailbox, the mail set is extracted to provide the mail set for the regular automatic update of the classifier. By using the latest mail sample set to train the classifiers, the classifiers can better determine the categories of new mail. 3, the design scheme of the front and background modules of the enterprise-oriented anti-spam gateway system is studied. The front desk includes system management, basic configuration, isolation area management, log management and other functional modules, background includes main control module, mail reception, original mail parsing, mail content analysis, phishing mail detection, mail forwarding, honeypot technology, In this paper, the design, implementation and test of each module of the system are described in detail, and the typical application scenarios are given. To sum up, this paper designs and implements an enterprise-oriented anti-spam gateway system. Through the research of the key technologies of the subject, not only from the point of view of improving the accuracy of spam filtering to improve the recognition ability of email filtering gateway, Moreover, an update mechanism of mail classifiers based on user feedback and honeypot technology is designed, which ensures the optimal recognition ability of the classifiers and improves the overall spam recognition rate of the gateway system through continuous learning and training. Through the design and research of this paper, we can filter the mail in and out of the enterprise in real time, reduce the bandwidth occupation of the enterprise network to a certain extent, ensure the stable operation of the mail server, and reduce the threat of spam and phishing mail to the enterprise network.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.098

【相似文献】