Title: TEMPORAL ISOLATION IN MIXED-CRITICALITY ENVIRONMENTS

ABSTRACT: Due to cost and related considerations, there is an increasing tendency in safety-critical embedded systems towards implementing multiple functionalities on a single shared platform. Certain features have been identified as being common to a large number of important emergent application domains of this kind; these features must therefore be taken into consideration in designing resource-allocation and scheduling policies for such integrated platforms. These features include the following:

1. Different applications sharing the same platform may have different criticalities, in both the sense that their contribution to the overall platform-wide mission is different, and that they may be subject to different certification requirements by statutory certification authorities;

2. These applications are typically implemented as collections of event-driven code each of which is embedded within an infinite loop (and hence essentially runs "for ever"); and

3. Often, the different applications share additional resources (other than CPU's); some of these resources are serially reusable rather than preemptive.

This combination of features gives rise to a very rich workload model, and some interesting scheduling problems that are not satisfactorily addressed using techniques from conventional scheduling theory. In this talk, I will describe a series of formalisms that have been proposed to represent workloads possessing these features, and describe several open resource-allocation and scheduling problems concerning such workloads.

BIO: Sanjoy Baruah is a professor in the Department of Computer Science at the University of North Carolina at Chapel Hill. He received his Ph.D. from the University of Texas at Austin in 1993. His research and teaching interests are in scheduling theory, real-time and safety-critical system design, and resource-allocation and sharing in distributed computing environments.