Integration testing of heterotic systems

Abstract

Computational theory and practice generally focus on single-paradigm systems, but relatively little is known about how best to combine components based on radically different approaches (e.g. silicon chips and wetware) into a single coherent system. In particular, while testing strategies for single-technology artefacts are generally well developed, it is unclear at present how to perform integration testing on heterotic systems: can we develop a test-set generation strategy for checking whether specified behaviours emerge (and unwanted behaviours do not) when components based on radically different technologies are combined within a single system? In this paper, we describe an approach to modelling multi-technology heterotic systems using a general-purpose formal specification strategy based on Eilenberg's X-machine model of computation. We show how this approach can be used to represent disparate technologies within a single framework, and propose a strategy for using these formal models for automatic heterotic test-set generation. We illustrate our approach by showing how to derive a test set for a heterotic system combining an X-machine-based device with a cell-based P system (membrane system).

Heterotic computing: exploiting hybrid computational devices

“…Heterotic systems will require adaptations of conventional techniques and inventions of new approaches. Stannett & Gheorghe [14] describe how the well-established X-machine testing paradigm can be extended to cover a particular case of heterotic computing, that of a controlling state machine communicating with a second, unconventional, system acting as an 'oracle'. They discuss what further research is needed to extend this approach to full heterotic systems, with continuous and continual time evolutions.…”