Analysis and protection of interprocess communications in real-time systems
Geoffrey F. Carpenter, David J. Holding, A.M. Tyrrell · Journal of the Institution of Electronic and Radio Engineers · 1988
Software for many distributed real-time systems comprises discrete processes executing in parallel on spatially distinct computers which are connected together by a communications network. The correct operation of such a system relies on the integrity of interprocess communications. In this paper an overview is presented of the different types of communication transaction required in safety-critical and time-critical situations, and of the communication primitives available for the design of distributed systems. Emphasis is placed on their behaviour under fault conditions, and on the ability of ‘watch-dog’ mechanisms to recognize communication failures in various communication structures. Each communication structure is analysed using graphical modelling techniques and a state space description produced. The powerful, formal, and analytical techniques associated with the state space description are then used to develop a design method which will allow the designer to integrate watchdog mechanisms into proper fault-tolerant structures for software in distributed real-time systems.