Research Areas
Dynamical systems & control - optimization, robustness, stabilization, sampling, inverse problems

Research Description
Dynamic system modeling, analysis and control techniques have been applied routinely to electrical, mechanical, chemical, and aerospace systems. Yet, these techniques are not standard tools in the tool chest of the computer scientist.

Just the same, control algorithms are essential to the success of the Internet and are implemented on many levels (e.g., media access control, network route control, transport control, queue management, reliable multicast control, and quality of service). There are numerous examples of deployed control algorithms that have been designed in an ad-hoc manner and are not robust. Without effective control of the dynamical properties of networks, stability and performance problems affect the commercial viability of the Internet. These include congestion collapse, distributed denial of service, route flapping, and large end-to-end latency.

General purpose operating systems are often required to process information according to dynamical requirements (e.g., real-time I/O, processing, and display of information) without those requirements being specified explicitly. Without effective control of the dynamical properties of computer operating systems, operating systems exhibit stability and performance problems, e.g., unbounded jitter, priority inversion, and starvation. Real time operating systems solve some of these problems by reserving required resources for tasks, but this approach is generally quite conservative when tasks have variable resource requirements.

My research focuses on the application of dynamical systems and control techniques to the design of computer networks and operating systems - for problems in CPU resource allocation for real-rate processes, multimedia content distribution and processing, fairness of flow control in the computer network, and prevention and mitigation of DDoS attacks on network resources.

Applications of Research
Resource allocation and dynamic system problems in computer systems, communication networks, network security, and environmental systems

Recent Research Collaborations & Projects

1. Division of Computer-Communications Research, CISE, National Science Foundation (NSF), through Oregon Graduate Institute: "Collaborative Research: Progress-Based Resource Management Using Control," (with David C. Steere and Jonathan Walpole) Sept. 1, 2000-Aug. 31, 2003, $205,000
2. Division of Electrical Communication Systems, ENG, National Science Foundation, through Oregon State University: "Collaborative Research: Progress-Based Resource Management Using Control," (with David C. Steere and Jonathan Walpole) Sept. 1, 2000-Aug. 31, 2003, $150,000
3. Kelly Foundation: "College of Engineering ABET 2000 Taskforce," (with G. Reistad, S. Randhawa, K. Douglas, M. Koretsky) Summer 1998-Summer 1999, $40,000
4. NSF & Sandia National Laboratory: "Grants for Source Localization Research," (with R. R. Mohler and A. Khapalov (1996-August 1998). M. H. Shor and her students (M.E. Alpay & B. Aktan) worked on this project ($110,000.)