ITR: Proactive Self-Tuning Systems for Ubiquitous Computing

Carnegie Mellon University, Pittsburgh PA

Investigators

Abstract

A system called Aura will be designed, implemented, deployed and evaluated. Aura will exploit the abundance of computing resources made possible by Moore's Law to reduce demand on the scarcest resource, human attention. Aura is targeted for environments involving wireless communication, wearable or handheld computers, and smart spaces. The research spans every system level: hardware, operating system, applications and end users. Aura applies two broad concepts. First, is proactivity, or the ability of a layer to act in anticipation of requests by a higher layer. Second, is self-tuning: layers adapt by observing the demands and adjusting performance and resource usage to match. Both techniques contribute to lowering demand on human attention. These ideas will be evaluated through controlled experiments and live use by a small community. The metrics used in evalutation include both traditional computer systems metrics plus user-centric metrics. There are three major research thrusts. First, operating system and networking capabilities required by Aura-like systems are investigated. Second, how high-level user intent can be captured and exploited for proactivity is explored. Third, how contextual information can be used is examined. A unique aspect of the research is its tight integration with education since the Aura prototype will be used in graduate-level courses.