Invited paper: Network digital twins for optical networks

•Network Digital Twins are emerging as important components of automated operations systems. The role and function of NDTs is discussed, as well as their use in practice, particularly supporting planning operations but also provisioning and management operations;•Accurate transmission performance prediction, in arbitrary whole-of-network configuration and state scenarios, is the key capability of NDTs for optical systems;•Machine learning methods are shown to have proved useful in constructing optimized impairment and device performance models;•An optical NDT product implementation is described, and its use is demonstrated in generic planning use cases and in the resolution of otherwise intractable provisioning problems. Network Digital Twins (NDTs) are recognized as important components of network operations automation systems. As analytical utilities within such systems, NDTs provide data-driven information that is useful in many operations optimization applications. The nature and role of NDTs, as well as their generic use cases, are discussed in detail, reflecting contemporary industry work and understanding. The specific nature of NDTs for optical transmission networks – optical NDTs – as well as their specific use cases of interest, are reviewed. Optical NDT function, architecture and constituent models are discussed, largely through the lens of a particular – now product – implementation, with reference to alternatives and other industry sources. Performance, function and utility are validated and demonstrated, including operational use of available applications directly supported by an optical NDT. Continuing areas of research interest and industry work are commented.