Washington Systems and Networking group
University of Washington, Seattle
Our group researches a broad set of issues in networking with past projects in measurement, p2p, performance and network architecture. We collaborate extensively on projects extending into security, programming languages, and computer architecture domains.
A key research thrust motivating our group is trying to understand what censorship will look like in 10 years. We’re interested in building tools now that will extend beyond the current arms race, or conducting research that will inform future tool developers. Along these lines, there are three independent research projects being conducted at UW.
- Censorship circumvention systems: we are building an in-browser proxy system in collaboration with Google. The system, which is called uProxy, relies on peer-to-peer communications between browser instances running on end-hosts and on social network connections between the users managing the end-hosts in order to provide blocking resistant communications. An earlier instantiation of the system, called Unblock, provided multi-hop proxy connections that are provably resilient to Sybil attacks.
- Censorship measurements: a second area that we are focussed on is deepening our understanding of how censorship is performed by different countries and networks. This work builds on our prior expertise with network measurements (including iPlane, an architectural model of the Internet, and reverse traceroute, which characterizes the harder-to-measure reverse path of conversations). Projects measuring connectivity blocking, performance degradation, or surveillance would be particularly interesting.
- Clean slate secure network architecture: we are also interested in designing a clean slate network architecture that is resilient to various byzantine attacks and provides security and privacy guarantees by construction. The goal of this project is to lay out a vision what a future rearchitecting of the Internet architecture should be in order to achieve network neutrality, censorship resistance, privacy, and security in the face of compromised routers. A key observation is that this can be achieved only by designing a network where the principle of least privilege is used all throughout the network stack.