A General Framework for Hard-Real-Time Application System Design

Principal Investigator: Sanjoy K. Baruah
Funding Agency: National Science Foundation
Agency Number: CCR-9972105

Abstract
The rate-monotonic analysis (RMA) methodology is a powerful and popular formal methodology for the design, analysis, and implementation of hard-real-time computer application systems. This project proposes to address certain shortcomings of RMA--its restrictive task model, and the limited options available in the methodology for overcoming resource-sharing bottlenecks. More specifically, the goals of this project will be:

• designing new abstract models of real-time tasks that may more accurately reflect important properties of application systems, and identifying rules for mapping application systems onto the most appropriate models.
• analyzing various task models in order to enhance our understanding of what intrinsic properties render a model infeasible from a static-priority feasibility analysis point of view, and identifying very general classes of tractable models.
• determining methods (algorithms/ heuristics) for assigning appropriate priorities to tasks specified in these general models.
• designing algorithms for overcoming resource-sharing bottlenecks, by obtaining a deeper understanding of the nature of typical shared resources and by exploiting their unique characteristics, and
• building software that both validates the usefulness of the theoretical results, and makes these results available to the system designer.