Dependable Systems of Systems

As systems necessarily become more integrated and increasingly complex through market demands for more features, technical risks and therefore business risks increase. It becomes correspondingly harder to show that the properties desired of these Systems of Systems (SoS) actually hold under normal or abnormal operation. In particular, it is hard to detect emergent properties of a SoS because properties of individual systems are not necessarily compositional, especially during failure. This paper describes the objectives of a project addressing the problem of Dependable System of Systems and other related research in the field of Automotive Electronics.

The capability being developed is based upon the scalable ‘Assumption-Commitment’[1] paradigm so that it can be applied to large and complex systems of systems. Through the use of a Formal description of the specification of each of the systems, we can explore the SoS to show weaknesses in terms of dependability and we have shown that the technique is scalable. We can therefore look at the way in which knowledge of individual systems (eg Black box, White box and shades in between) can be used to mitigate against failure of the SoS. We show how this is of use to system integrators such as those in the automotive or aerospace sectors. The value to the integrator is because they are vulnerable through an incomplete understanding of supplier's equipment and the suppliers are similarly unsighted on how the system integrator intends to exploit their equipment. In doing so, we enable the integrator to explore the SoS before commitment to an architecture and individual system suppliers. The technology exposes and expresses risks in a timely and understandable manner, thus enabling business decisions to be well informed and relevant risks to be mitigated.