Task-oriented modelling of autonomous decentralised systems

An ongoing project of the Programme for Highly Dependable Systems (PHDS) at the University of the Witwatersrand is the development of a dependable decentralised system using readily available hardware and software components. An experimental system has been developed to support multiple-task applications with different levels of criticality. Fault-tolerant protocols are used to detect faults, to mask incorrect results from faulty nodes. A task in a faulty node can he recovered through a system reconfiguration or task reallocation. A faulty node can he repaired and reintegrated into the system. This paper focuses on modelling the system under the occurrence of faults, reconfiguration and repair. The method developed can be used to evaluate individual task's reliability, risk and availability.