A distributed real-time operating system

Multiple processor systems with shared memory are used in embedded real-time systems for various applications. The authors present an approach to the development of operating systems for multiple processor systems with shared memory, assuming that memory is used for data exchange only, not for code storing. Compared to alternative approaches, this one enables faster and cheaper development of reliable operating systems. The approach is based on: the implementation a commercial uniprocessor operating system kernel together with the development of: a global kernel, a local and a remote IO system; minimization of the global kernel providing services sufficient for many embedded multiple processor systems applications; a new model for interprocessor communication, that helps achieve satisfying performances of the distributed operating system; an optimally structured IO system, for both local and remote IO. This approach has been used in the design and implementation of a real-time signal processing system.