Object-oriented and functional software design for distributed real-time systems

Real-time systems in applications like command, control, communications and intelligence require complex distributed systems with many interacting software components, heterogeneous processing systems and sharing resources. These systems should satisfy not only the functional requirements of application software, but also the specified timing constraints on the execution of the software, despite faults and failures. In addition, parallelism needs to be expressed in the design of such systems and exploited on the target distributed computing systems. In this paper, an approach to software design for distributed real-time computing systems, based on the PROOF computation model which integrates object-oriented and functional paradigms, is presented. To support adaptability of the software system to a predictably changing environment, our approach supports multi-versions of a method definition, synchronous communication within objects, asynchronous communication among objects, encapsulation of timing constraints in objects and expressing parallelism in object-level and method-level. Our design approach consists of the high-level object-oriented design and object design phases. Our design approach is illustrated with a hypothetical chemical plant simulation system.