Deriving the optimal structure of N-version software under resource requirements and cost/timing constraints

Abstract

The paper presents an approach to systematically derive the optimal structure of fault-tolerant software for safety-critical control processes, with respect to resource requirements and cost- and timing constraints. For the modelling of the control process, a graph model description is used; basic static distributed scheduling and allocation decisions for the tasks are performed and task computation times are defined. Adaptive optimization techniques are used to derive the optimal solution for the given requirements. The application of the method in practice is illustrated by the example of a software system for spacecraft control.