Electronic Systems Design Seminar
http://www-cad.eecs.berkeley.edu/esd-seminar


 

Dynamic Power and Fault-Tolerance Management
for Ambient Intelligent Systems

Prof. Diana Marculescu
Department of Electrical and Computer Engineering
Carnegie Mellon University

Monday, November 17th, 2003, 4 - 5 pm
540A/B Cory Hall (D.O.P. Center Classroom)

Abstract

Ambient Intelligent Systems provide an unexplored hardware platform for executing distributed applications under strict energy constraints. These systems must respond quickly to changes in user behavior or environmental conditions and must provide high availability and fault-tolerance under given quality constraints. Techniques to provide reliable computation in the presence of failures must do so while maintaining high performance, with an eye toward energy efficiency. When possible, they should maximize battery lifetime in the face of battery discharge nonlinearities.

This talk introduces the concept of adaptive fault-tolerance management for failure-prone, energy limited ambient intelligent systems and proposes two approaches for prolonging application lifetime, under stringent energy constraints. First, pre-copying with remote execution is proposed as a novel, alternative technique of code migration to enhance system lifetime for ambient intelligent systems. In this case, self-management of the system is considered in two different scenarios: applications that tolerate graceful quality degradation and applications with single-point failures. Second, a classification of local decision policies for achieving system-wide reliability is proposed. In order to judge the efficacy of the proposed algorithms for dynamic fault-tolerance management, a set of metrics, for characterizing system behavior in terms of energy efficiency, reliability, computation performance and battery lifetime, is presented. Analytical results show up to 80% increase in application lifetime for various levels of redundancy. For an example platform employed in a realistic evaluation scenario, it is shown that system configurations with the best performance or lifetime are not necessarily those with the best combination of performance, reliability, battery lifetime and average power consumption.

Speaker

Diana Marculescu is currently an Assistant Professor of Electrical and Computer Engineering at Carnegie Mellon University. She received her M.S. in Computer Science from "Politehnica" University of Bucharest, Romania in 1991 and her Ph.D. in Computer Engineering from University of Southern California in 1998. Her research interests include energy aware computing, CAD tools for low power systems and emerging technologies (such as electronic textiles or ambient intelligent systems). She is the recipient of a National Science Foundation Faculty Early Career Development Award (2000-2004) and of an ACM-SIGDA Technical Leadership Award (2001-2003). Diana Marculescu is a member of Executive Board of the ACM Special Interest Group on Design Automation (SIGDA) and a member of IEEE Circuits and Systems (CAS).



Contact 
©2002-2018 U.C. Regents