基于SIP的SPIT防御方案的设计与实现

发布时间：2018-10-17 14:10

【摘要】：摘要：伴随着IP多媒体子系统(IP Multimedia Subsystem, IMS)的引入,分析报告预言网络电话(Voice over Internet Protocol, VoIP)用户数必定增长。相应地,在未来几年,网络电话垃圾(Spam over Internet Telephony, SPIT)将会成为一个非常严重的问题。 SPIT,是指通过VoIP网络传送的大量未经请求的呼叫,通过播放实时媒体信息向被叫进行商品推销、商业广告推广或者恶意呼叫,对正常用户生活造成严重困扰,严重浪费网络资源。使公共服务部门无法向公众提供及时有效的服务。本论文的主要工作如下：首先对IETF (Internet Engineering Task Force)等组织提出的反垃圾电话技术进行分析,不难发现,尽管目前已经提出了大量的SPIT防御方法,但相关的研究仍然处于一种初始阶段。然后,本文设计了一种更有效的SPIT防御方案,是基于会话发起协议(Session Initiation Protocol, SIP)的,因为SIP解决方案占据了当前企业市场的主体。该方案由三种关键机制构成,既可以检测机器发起型SPIT,也可以过滤人工发起型SPIT,都是有效的。第一种机制基于图灵测试原理,使用一个判别系统解决SPIT研究领域的“首次呼叫”问题。第二种机制是利用概率神经网络(Probabilistic Neural Network, PNN),设计了一个SPIT分类器。本文的第三种机制叫做蜜机(Honeyphone),是受蜜罐(Honeypot)思想的启发而想到的。最后,对三种机制进行了实现,测试和分析结果显示,整体方案在完备性、实时性和可用性方面性能良好。同时,还归纳总结了现有技术方案的优缺点。总的来说,本文方案是有效的,这三种机制能够各自实现独立的功能,也能够部署在不同位置,本文建议把这三种机制融合,将得到的信息结合在一起,以使决策更加准确。
[Abstract]:Absrtact: with the introduction of IP multimedia subsystem (IP Multimedia Subsystem, IMS), the report predicts that the number of (Voice over Internet Protocol, VoIP) users will increase. Accordingly, in the next few years, (Spam over Internet Telephony, SPIT) will become a very serious problem. SPIT, refers to a large number of unsolicited calls over the VoIP network. By playing real-time media information to be called to sell goods, commercial advertising or malicious calls, the normal user life caused serious problems, serious waste of network resources. Make it impossible for public service departments to provide timely and effective services to the public. The main work of this thesis is as follows: firstly, the anti-spam telephone technology proposed by IETF (Internet Engineering Task Force) and other organizations is analyzed. It is not difficult to find that although a large number of SPIT defense methods have been proposed, the related research is still in an initial stage. Then, this paper designs a more effective SPIT defense scheme, which is based on the session initiation protocol (Session Initiation Protocol, SIP), because the SIP solution dominates the current enterprise market. The scheme consists of three key mechanisms, which can detect both machine initiated SPIT, and filter manual initiated SPIT,. The first mechanism is based on the Turing test principle and uses a discriminant system to solve the "first call" problem in the field of SPIT research. The second mechanism is to design a SPIT classifier using probabilistic neural network (Probabilistic Neural Network, PNN),). The third mechanism in this paper, called Honey Machine (Honeyphone), is inspired by honeypot (Honeypot). Finally, three mechanisms are implemented, and the results of test and analysis show that the overall scheme has good performance in completeness, real-time and usability. At the same time, the advantages and disadvantages of the existing technical scheme are summarized. In general, the proposed scheme is effective. These three mechanisms can achieve independent functions and can also be deployed in different locations. This paper proposes to combine the three mechanisms and combine the information to make the decision more accurate.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TN916.2

【参考文献】