On the meaning of modes in uniprocessor real-time systems

One way to manage the complexity of the design and the implementation of large and adaptive real-time systems is to partition the design into modes of operation. This approach allows designers to break down the complexity of these systems into a number of layers of control (modes), which are then multiplexed in time (by means of mode changes) during system operation. However, due to a lack of a uniform view of modes within the literature, the task of successfully designing complex real-time systems with modes of operation becomes more difficult, if not compromised [1]. In this paper we consider the following recurring issues of modes in real-time systems, both from the application and design concerns: 1) What is a mode of operation, i.e. which features are exclusive to a particular mode of operation that allow its identification and distinction from other modes? 2) What can be understood by the concept of mode changes in realtime systems? The examination of these issues is lacking in the literature and will facilitate the analysis, design and construction of the next generation of complex real-time systems.