A modeling methodology for real-time/multimedia operating systems

Abstract

Real-time scheduling theory attempts to guarantee that a real-time task set always meets its deadlines. Historically, there has existed a wide gap between real-time scheduling theory and the reality of applying the theory to task sets implemented via real-time operating systems (RTOSs). The paper provides a framework to account for implementation costs in real-time scheduling theory. In addition, an engineering methodology that allows users and developers to accurately model and evaluate RTOSs is presented. We use this methodology to model three different commercial real-time operating systems that are being used in multimedia applications. We show how to use the RTOS scheduling models to evaluate the performance and design of RTOSs.