The infrastructure upon which the functioning of society depends is composed of complex ecosystems of systems. Consequently, we must reason about the properties of such ecosystems, which requires that we construct models of them. There are very many approaches to systems modelling, typically building on complex structural and dynamic frameworks. For example, there are simulation modelling tools based on a theoretical treatment of a ‘distributed systems metaphor’ for system modelling. These tools are based on the concepts of location, resource, and process. They are well developed and have proved valuable in a range of settings. Arguably, however, the foundations of modelling technologies remain underdeveloped. Our purpose here is to explore a foundational modelling framework based on minimal assumptions, starting from a primitive notion of behaviour, and to show that such an approach allows the recovery of the key ideas, including a generalized CAP theorem, required for effective modelling of and reasoning about ecosystems of systems. We establish a logic of behaviours and use it to express local reasoning principles for the compositional structure of systems.