Issues in schedulability analysis of real-time systems

Abstract

Due to the critical nature of real-time systems, there is an ever growing burden on the designer to not only guarantee that the tasks would meet their deadlines at design time, but would continue to do so as the system evolves. This implies that the schedulability analysis has to he robust. We identify the sensitivity of schedulability analysts to the task execution times. The impact this parameter has on the schedulability analyses is captured by formulating a general scaling problem. This problem is shown to relate to the problems of scalability, portability and execution time estimation. A technique to solve this general problem is developed. A proof of correctness and optimality of the technique is presented.