A Prevention System for Spam over Internet Telephony

Internet Telephony is realized by way of instant messaging and the transmitted data are voice rather than text packets, therefore it is more difficult to identify and/or prevent Spam over Internet Telephony (SPIT) than SPAM  junk e-mail. With the rising popularity of Voice over IP (VOIP), SPIT becomes a serious issue on security nowadays. This paper proposed an effective SPIT defense system, which can automatically collect the characteristics of an Internet Telephony and perform k-Nearest Neighbor classification to determine whether it is a SPIT. If convicted of being SPIT, the proposed system will immediately update the built-in blacklist in the SIP server to prevent SPIT from the same sender subsequently. The call connection protocol adopted in this study was Session Initiation Protocol (SIP). In the experiments, malicious software named Spitter was used to generate SPIT. The experimental results suggest that the system can get the overall accuracy up to 93.65%.