ct-er-collaborative-research-detecting-spam-in-ip-multimedia-communication-services

VoIP, along with text-based systems such as instant messaging, is likely to gradually replace the circuit-switched telephone system. VoIP offers the promise of global and low-cost communications. However, these attributes also make VoIP a promising target purveyors of unsolicited bulk communications. Unsolicited phone calls, sometimes referred to as "spit" (Spam over Internet Telephony) can be even more annoying and disruptive than email spam. For example, such phone calls may arrive in the middle of the night or may disrupt a business meeting. Such calls may be used for advertising or for soliciting private information by social engineering, extending the threat of phishing into a new realm.

The social meaning of trust, reputation, and nuisance are quite different for interactive voice communication compared to e-mails.

This proposal describes a comprehensive call admission control framework for accepting, rejecting or forwarding incoming calls to a voice mail system. We will be formalizing notions of trust and reputation for voice communications. One such mechanism is the notion of trust paths that reflect previous communication relationships between users and domains. We plan to devise methods for integrating several social-technical methods for minimizing false rejections.

The methods will be tested in a VoIP security testbed being developed by four universities (UNT, Columbia, Purdue and UC Davis). Results will be fed back into on-going standardization activities in the IETF and elsewhere.