Graduation Semester and Year




Document Type


Degree Name

Master of Science in Computer Science


Computer Science and Engineering

First Advisor

Matthew Wright


Voice over IP (VoIP) is one of the fastest growing telecommunicaiton market today. However with great popularity comes greater security risk. When we communicate over the Internet we expose our identity to all the prying eyes, even if we do not want to. Even though VoIP data can be secured by using encryption, sometimes the relationships between users can be sensitive. To provide stronger privacy for these users, we propose anonymous VoIP (aVoIP) - a means to secure user privacy in VoIP. The biggest challenge in introducing anonymity to VoIP system is in providing acceptable Quality of Service (QoS). Users expect a voice quality comparable to that of traditional telephone network. The combination of affordability, convenience and good voice quality has fueled significant development in VoIP system nowadays. However things go little awry with introduction of anonymity in the picture. To achieve anonymity we may need to employ cryptographic operations, random re-routing and traffic manipulation - all of which might add some extra delays in VoIP operation. For aVoIP to be feasible, we need it to be reliable, scalable and provide acceptable QoS. Rather than reinventing a wheel, we looked for existing anonymity systems currently deployed over the Internet to base the architecture of aVoIP on. With hundreds of thousands of users all over the world, Tor has proved to be a reliable, scalable and stable anonymity system. Hence we foresee aVoIP to have architecture similar to Tor. However, Tor has been designed to provide low-latency anonymity only to TCP-based applications, whereas most of the real-time and multimedia systems need fast and efficient connection-less protocol like UDP. So in our quest to determine the feasibility of aVoIP over current Internet infrastructure, we worked towards modifying the Tor system to support UDP. We studied the Tor system in depth, analyzed the complications raised while supporting UDP, proposed a new design and implemented it. We have also developed a test environment to test our design over the PlanetLab nodes.


Computer Sciences | Physical Sciences and Mathematics


Degree granted by The University of Texas at Arlington