ABSTRACT
Often, people say that real-time systems must react fast to external events. Such a definition, however, is not precise, because processing speed does not provide any information on the actual capability of the system to react timely to events. In fact, the effect of controller actions in a system can only be evaluated when considering the dynamic characteristics of the controlled environment. A more precise definition would say that a real-time system is a system in which performance depends not only on the correctness of the single controller actions, but also on the time at which actions are produced [1]. The main difference between a real-time task and a nonreal-time task is that a real-time task must complete within a given deadline. In other words, a deadline is the maximum time allowed for a computational process to finish its execution. In real-time applications, a result produced after its deadline is not only late, but can be dangerous. Depending on the consequences caused by a missed deadline, real-time activities can be classified into hard and soft tasks [2]. A real-time task is said to be hard if missing a deadline may have catastrophic consequences in the controlled system. A real-time task is said to be soft if missing a deadline
ZURA: “2824_C012” — 2005/6/21 — 20:02 — page 2 — #2
causes a performance degradation, but does not jeopardize correct system behavior. An operating system able to manage hard tasks is called a hard real-time system [3,4].
