Abstract
Current software systems are increasing in scale, distribution, and degree of interaction with their environment. Massively distributed embedded systems and multi-stage networked server clusters are two examples. Ensuring performance guarantees in these systems is an important research topic with many practical applications. In this talk, I will present some of my recent work which focuses on efficient analysis and control of temporal behavior in embedded and networked computing systems. First, I will introduce Universal Feasible Region Analysis, a new theoretical schedulability analysis framework for ensuring system schedulability under general workload. Its advantages are scalability for large networks and applicability to arbitrary, fixed-priority scheduling policies. I will illustrate the use of this framework for efficient admission control in networked embedded systems. Then I will present Queueing Model Based Feedback Control, a novel performance management framework for networked server systems. It integrates two separately developed theories--feedback control and queueing theory--into a single analytical framework. Hence it can provide better performance guarantees under dynamic workload compared with previous proposals.

Speaker Bio
Xue Liu defended his Ph.D. thesis in Computer Science in 2006 at the University of Illinois at Urbana-Champaign, where he is advised by Professor Lui Sha. He will join School of Computer Science at McGill
University in Montreal, Canada as a tenure-track assistant professor. Xue got his B.S. in Mathematics and M.S. in Automatic Control both from Tsinghua University, China. His research interests are in real-time and
embedded systems, networked server performance management, and software reliability. He received the Mavis Memorial Fund Scholarship Award from College of Engineering of UIUC in 2005 for excellent academic
performance, research accomplishments, and demonstrated leadership in engineering education, and the C. W. Gear Outstanding Graduate Award in 2002 for being the best graduate student from the UIUC Department of Computer Science. He was also recipient of the prestigious Saburo Muroga Fellowship and Ray Ozzie Fellowship.