Performance engineering of Cyber-Physical Systems

Cyber-Physical Systems (CPS) are the consequence of today’s technology progression. They were born as evolution of a very large family of systems called embedded systems, complex systems and system of systems and many others. Technology improvements are computational power, more powerful telecommunication architectures and more efficient software and algorithms that characterize the smart features that are so widespread and used since few years ago. CPS are complex systems and the evaluation of their performance is a crucial task. Waiting the implementation to assess the system performance is quite costly, whereas at the development stage, refactoring actions might have lower costs. Moreover, CPS, due to their nature, have to work satisfying different types of requirements, e.g., safety-critical or real-time constraints. When the system is complex and refactoring actions are costly when performed at the implementation stage, the capability of evaluate system performance in the early stage of the development process is very important. A large set of challenges in many activity of the engineering and design process have been raised by CPS. The communication and collaboration features, among system agents and components, have made it difficult to apply the well-established design and performance evaluation methodologies of the previous systems generations, such as embedded systems. In literature, there are formalisms and tools specialised in modelling and evaluating software, hardware or physical components and also communication infrastructures and protocols and so forth. However, a modelling formalism for software has difficulties in modelling hardware features such as laws of physics and a simulation tool for software struggle in accurately simulate also mechanical, hydraulic or electric dynamics. Vice versa the situation is very similar, formalisms and tools for hardware and physical components are not capable, or very little, of modelling and evaluating software elements. This doctoral work aims to investigate new possibilities about these research challenges. In particular, the main topic of this research work is to present a novel methodology to performance engineering Cyber-Physical Systems, exploiting and leveraging, if possible, state of the art ideas and concepts but also well-established formalisms, tools, and techniques. The main goals of this work are to be a step forward towards the innovation of engineering CPS and to propose a novel methodology that would help to establish future innovative performance engineering and model-based approaches for CPS.