Digital Twins are nowadays gaining momentum in smart manufacturing context. They are generally intended as virtual replicas of physical systems, capable to run on different simulation disciplines that synchronize the virtual and real systems, thanks to sensed data and connected smart devices, mathematical models and real time data elaboration. More specifically, they blend the virtual models and capabilities of Cyber-Physical Systems (CPS) (i.e. computing, communication, data collection, intelligent sensoring), with the potential of big data analysis and the pervasive connection allowed by Industrial Internet-of-Things, to have an advanced real-time perception of the physical assets and a prompt prediction of their behaviour. Digital Twins promise high benefits when used to support the design, management, monitoring, control, performance evaluation and optimization of manufacturing assets (i.e. manufactured products, production equipment and production systems). Moreover, they open the way to extensive simulations under various scenarios and conditions, with the possibility to be up-dated in real-time, to detect anomalies and to conduct accurate diagnostics and prognostics. Eventually, they open new possibilities to make a joint production and maintenance scheduling, relevant to assure more robust performances in stochastic environments, as well as to assess and optimize performances at different levels, from components and machines to production systems.