Andrew Mao

The design and implementation of the A3 application-aware anonymity platform

This paper presents the design and implementation of Application-Aware Anonymity (A^3), an extensible platform for rapidly prototyping and evaluating anonymity protocols on the Internet. A^3 supports the development of highly tunable anonymous protocols that enable applications to tailor their anonymity properties and performance characteristics according to specific communication requirements. To support flexible path construction, A^3 uses a declarative language to compactly specify path selection and instantiation policies. We demonstrate that our declarative language is sufficiently expressive to encode novel multi-metric performance constraints as well as existing relay selection algorithms employed by Tor and other anonymity systems, using only a few lines of concise code. We experimentally evaluate A^3 using a combination of trace-driven simulations and a deployment on PlanetLab, as well as a case-study of A^3-enabled voice-over-IP communication. Our experimental results demonstrate that A^3 can flexibly and efficiently support a wide range of path selection and instantiation strategies at low performance overhead.