Digital twin technology has emerged as a powerful paradigm for modelling, simulating, and managing complex systems virtually across their lifecycle. However, as digitalisation efforts grow exponentially and the endeavours to digitise more objects and assets continue at a rapid pace, more than a single digital twin may be required to give holistic insights and provide multiple decision-making mechanisms in addition to other services. So, it is becoming increasingly necessary to connect and couple various digital twins across different systems/domains so that each can offer a specific service to the whole digital twin. In this paper, we explore the Digital Twins Federation (DTF) concept, which is envisioned as a unified, integrated, interconnected, interoperable, and evolving ecosystem that couples and comprises a group of digital twins. The concept can be utilised in a broad context and targets large-scale systems. The paper discusses the modelling and (re-)engineering of digital twins towards federation. Moreover, examples of architecture and abstract models for the multi-robot warehouse case, entailing the key details of the concept, are provided. Outlining its enabling technologies, current challenges and barriers, the paper suggests a plan and a roadmap for the next steps to realise the vision and enable integration, collaboration, and coordination across an ecosystem of federated digital twins.