Spam over internet telephony: Prototype implementation

This paper presents a prototype implementation of the reputation-based spam detection scheme, first presented in [4]. Different components were developed, in particular the Reputation Manager (RM) that collects the appropriate information from the contact lists of the SIP subscribers and builds a social network. The latter provides the possible paths that might exist between any two SIP subscribers. This social network is endowed with a metric that allows to assign reputation values to the mentioned paths. Based on these scored paths, a decision is taken for accepting or rejecting calls. The result of our implementation is a working mechanism that tolerates zero false acceptance and an average of 4% false rejection. Our work demonstrates the scalability of the suggested solution as well as the feasibility of using the reputatioon concept for detecting SIP spam.