Design of Real‐Time Systems

摘要

Abstract Real‐time computer systems are characterized by the existence of timing constraints as well as logical correctness requirements on computations they carry out. The timing constraints are statically determined prior to system operation time from the characteristics of physical systems they interact with. In hard real‐time systems , a timing failure is considered catastrophic and a guarantee should be given prior to runtime that every timing constraint will be satisfied throughout the system operation time. Examples are found in application domains such as avionics, process control, automated manufacturing, and robotics. In this article, the issues and approaches on real‐time system design addressed in the uniprocessor environment. In addition to the timing constraints, the characteristics and arrival patterns of real‐time tasks should be known ahead of system operation time. This information must be used in system design time to guarantee that all the timing constraints will be satisfied. These article is organized as follows. In the next section the real‐time scheduling problem to be studied in this article is formally defined. We tried to come up with a problem definition that is as general as possible in terms of the types of constraints and requirements that may be specified. Then priority/time‐based scheduling mechanisms are briefly explained and compared in terms of their capability to satisfy various system requirements. Fixed‐priority based scheduling scheme is presented mainly because it has been widely studied and adopted in many real‐time systems, and because similar results are available for other priority based scheduling schemes such as the earliest deadline first (EDF) scheduling algorithm. Then dynamic time‐based scheduling scheme is briefly summarized. A conclusion of the article follows.