TC: Small: Provably Anonymous Networking Through Secure Function Evaluation

Georgia Tech Research Corporation, Atlanta GA

Investigators

Abstract

Anonymous communication in computer networks generally relies upon the filtering of traffic through a cascade of mixes. Clients select a number of proxy nodes that modify and obscure the origin of a message so as to make the detection of a relationship between source and destination extremely difficult. While this architecture is known to not be robust against a globally passive adversary, an ever growing body of literature has demonstrated that even moderately capable adversaries can link the communications between two parties using these networks. Accordingly, new techniques ensuring stronger levels of anonymity must be explored. This project will develop a new architecture for anonymity networks offering cryptographic guarantees of anonymity based on a foundation of Secure Function Evaluation. These Provably Anonymous Networks (PANs) protect the participants of communication from traffic analysis attacks by remaining unaware that the exchange of messages has occurred. However, realizing a more trustworthy architecture is not simply the result of haphazardly assembling the above components. Rather, this system must be carefully composed so as to avoid the leakage of information useful in the identification of a communication channel. The results of this project will not only be used to enhance graduate and undergraduate curriculum, but will also be used to develop a tool to assist members of the Carter Center administer observe elections without fear of eavesdropping.