The vision of the smart factory with its interconnected systems is based on the seamless (real-time) integration of data across the information system (IS) landscape. Yet, due to the existence of many legacy systems, this task is far from trivial. In industrial practice, the IS landscape comprises systems of different application functionality which can be characterized as technologically heterogeneous, e.g. transaction processing versus real-time systems. Integrating such systems has always been a major challenge and constraining force for many organizations. This problem is receiving renewed attention in the context of the implementation of the digital twin in manufacturing. Due to its central role in the IS landscape, the digital twin needs to communicate with a number of heterogeneous applications to achieve its full potential, i.e. achieving a complete virtual representation of an asset, process or product. This research analyses the integration requirements from the perspective of the digital twins’ application functionality. In particular, we provide an explicit mapping of the integrations needed between the digital twin and existing information systems (IS) in manufacturing, which serves as a basis to better understand integration issues. These findings provide an explanation for and a conceptualization of some of the challenges that emerge when transforming towards an interconnected smart factory.