Detecting More SIP Attacks on VoIP Services by Combining Rule Matching and State Transition Models

The Session Initiation Protocol (SIP) has been used widely for Voice over IP (VoIP) service because of its potential advantages, economical efficiency and call setup simplicity. However, SIP-based VoIP service basically has two main security issues, malformed SIP message attack and SIP flooding attack. In this paper, we propose a novel mechanism for SIP-based VoIP system utilizing rule matching algorithm and state transition models. It detects not only two main attacks, but also covers more SIP attacks. Instead of simply combining rule comparison and counting number of SIP messages, we develop secure RFC 3261 rules based on existing RFC 3261 rules, so that proposed mechanism shows 26% higher detection rate for malformed attack. Moreover, we utilize session information and define the features of each state in order to detect abnormal situations including SIP flooding. As the result, it is shown that the proposed mechanism provides not only higher accuracy, but also covering more SIP attacks including two main attacks.