Increasing feature and performance demands in today's systems, IoT, Cyber-physical systems (CPS), etc. leads to increasingly complex components, systems and systems of systems. Designing, validating, integrating and verifying that they deliver the expected features within performance, energy, temperature, robustness, safety and security bounds becomes unaffordable in time and cost. Hence, new paradigms are needed for the design of components, systems and systems of systems so that:

• Design and integration are eased by allowing lower dependence on the internal state of components so that only external behaviour delivers enough information.
• Modelling is simplified so that the external behaviour (measurements) delivers most of the information needed, thus removing the need for detailed models of complex components, systems, etc.
• Features such as performance, energy, temperature, robustness, safety and security are easy to predict with probabilistic and statistical means without needing detailed models.

All these properties together will allow the design, validation, integration and verification of increasingly complex components, systems, systems of systems and CPS, as well as building a more powerful IoT without incurring higher costs and keeping time-to-market low.