Armored twins: Flexible privacy protection for digital twins through conditional proxy re-encryption and multi-party computation

Digital twins, i.e., up-to-date digital copies of a physical object maintained in the cloud, make it possible to conveniently review a physical object’s state, indirectly interact with the physical object, or perform computations on the object’s state and history – also in combination with data from other digital twins. The concept of digital twins has seen wide uptake in Internet of Things use cases, e.g., in manufacturing to monitor a product’s lifecycle, or precision medicine to provide personalized treatment. Besides these benefits, challenges arise, especially if the involved data producers, clouds and data consumers are not in the same trusted domain: Who owns and controls the data? Are the parties (e.g., cloud) sufficiently trusted to handle privacy-sensitive data? In this work, we propose ARMOREDTWINS, i.e., a system for digital twins that protects the confidentiality of digital twin data while providing flexible and fine-grained sharing by employing key-policy conditional proxy re-encryption to enable processing on subsets of the data. Alternatively, to support computation on very sensitive data, our system integrates secure multi-party computation, which does not reveal the data items to the individual nodes performing the computation. Benchmarks of our implementation highlight the system’s feasibility and practical performance.