Privacy in networked systems goes beyond the secrecy of exchanged data; it is equally critical to protect the identities of communicating parties and the paths of data flow. Availability of such networking information is not only a violation of user privacy, but also provides a convenient platform for a malicious adversary to launch more powerful attacks. Anonymous network systems are networks of special routers and proxy servers that provide protection using cryptographic tools and covert relaying. Commonly deployed systems are, however, vulnerable to timing analysis of packet transmissions and do not provide the necessary protection under resource limitations.
The principal investigators will investigate the provably optimal design of anonymous network systems under constraints on resources and desired network quality of service (QoS). They will develop a novel quantitative measure of anonymity that takes into account the complete information available to an adversary and the resource and topological constraints of a network. Using the quantitative model, they will explore the following three critical problems in network system design: optimal resource allocation, stochastic routing, and topological design. If successful, the analysis will give rise to novel networking algorithms to achieve anonymity in information exchange that cater to modern paradigms of network design such as mobility and application independence.
Broader Impact: The research will have a significant impact in delivering provable privacy for networked systems. The research promotes graduate and undergraduate education in network security and provides access to recent advances in the field to community college students and law enforcement personnel through distance learning programs. In addition, awareness of security vulnerabilities in social networks will be broadened for middle to high school students.