Reputation Based SPIT Detection And Defense In VoIP

With the development of VoIP(Voice over IP) technology and VoIP market, a technology called SPIT(Spam over Internet Telephony) which is a voice based spam attack will become a serious security problem to VoIP transaction, it's serious degree can even compare to spam, so it is necessary to consider the way to solve this problem.Reputation module is a kind of module that can find out the trust degree of users by analyzing the feedback information from other users after transaction. The module has been widely used in different kinds of internet transaction, and gains good effect. The thesis introduces it into VoIP system to solve SPIT.After analyzing and comparing several reputation modules which based on different theories, the thesis designs a VoIP based reputation module by combining characteristic of VoIP. Different from other common reputation module, this module is based on Dempster-Shafer theory, and includes subjective estimate and objective estimate that can make sure the accuracy of statistical result.In order to guarantee the storage security of reputation information, the thesis introduce a heterogeneous distributed network, the network can not only guarantee the storage security of reputation information but also keep the principle of call-processing in VoIP system. Moreover, the network offers a DHT(Distributed Hash Table ) algorithm through which the system can search router and reputation information quickly.At last, the thesis uses a simulation experiment which simulates the real environment of malicious voice attack to test the SPIT filter performance of the reputation module. Simulation results indicated that, the reputation module gets a high interception rate to voice spam, and a low mis-interception rate to non-voice spam, further more, the module can maintain its good performance when in a relatively poor environment.