PDL PEOPLE

Research Interests:

My primary interests are in the broad areas of network protocols and distributed network applications. In the past, I have worked on topics such as transport/routing protocol interactions with wireless networks, fast protocol stack implementations, RAID system design, performance prediction for Internet transfers, firewall design, and improvements to the TCP protocol. My current work includes overlay network design, Web server benchmarking, mobile computing/networking, and new approaches to congestion control. The following three projects are examples of my current research efforts.

Congestion Control. In 1999, I co-founded the Congestion Manager (CM) project with collaborators at MIT (http://inat.lcs.mit.edu/projects/cm/). This project led to the creation of the ECM IETF Working Group which standardized the use of the CM for end-point congestion management. We are currently exploring techniques to use the CM to allocate bandwidth among hosts in a network. An important goal of this work is to require no router support. Requiring router support has severely limited the deployment and use of existing techniques for bandwidth allocation.

Mobile Networking. A second project aims to design a new networking and operating system infrastructure for the next generation of ubiquitous computing applications. The project is especially interested in 1) improving the bandwidth and power efficiency of network protocols, 2) making it possible for applications and protocols to adapt to the availability of power, availability of bandwidth and other environmental conditions, 3) designing service discovery approaches that can handle the query types that are likely for mobile applications, and 4) making this environment support the easy introduction of new link technologies, network protocols and applications.

Wide-Area Distributed Applications. I am also currently working on developing better support for distributed wide-area Internet applications. While a few such applications exist today (e.g., Akamai, peer-to-peer systems), the bulk of Internet applications follow the client-server paradigm. The difficulty of designing, building, testing, and maintaining distributed applications has been a key factor preventing their deployment. This project is especially interested in developing tools, building blocks and network infrastructure aimed at making such applications easier to develop and deploy.