Real-time systems technology, traditionally developed for safety-critical systems, has recently been extended to support novel application domains, including multimedia systems, monitoring apparatuses, telecommunication networks, mobile robotics, virtual reality, and interactive computer games. Such systems are referred to as mft real-time systems, because they are often characterized by a highly dynamic behavior and flexible timing requirements. In such systems, missing a deadline does not cause catastrophic consequences on the environment, but only a performance degradation, often evaluated through some quality of service parameter.
Providing an appropriate support at the operating system level to such emerging applications is not trivial. In fact, whereas general purpose operating systems are not predictable enough for guaranteeing the required performance, the classical hard realtime design paradigm, based on worst-case assumptions and static resource allocation, would be too inefficient in this context, causing a waste of the available resources and increasing the overall system cost. For this reason, new methodologies have been investigated for achieving more flexibility in handling task sets with dynamic behavior, as well as higher efficiency in resource exploitation.