Unlocking Secure Industrial Collaborations through Privacy-Preserving Computation

value data science can have for industries. A prime example is the research cluster “Internet of Production” [L1], which aims to turn data into value throughout the entire product lifecycle, i.e., production, development, and usage. The cluster, which was established in 2019, brings together more than 200 engineers and computer scientists from more than 35 institutes at RWTH Aachen University and the Fraunhofer Society. Its key vision is to interconnect companies with the aim of exchanging knowledge and know-how globally (Figure 1), i.e., advancing use cases within and across domains to establish reliable, cost-efficient, sustainable, and accountable production. Not surprisingly, the involved industrial stakeholders mandate strict confidentiality concerning their data as they fear a loss of control [1]. To address these concerns, privacy-preserving computation with its diverse building blocks, such as homomorphic encryption (HE), private set intersection (PSI), or oblivious transfers (OTs), can act as a key enabler. Here, industrial settings provide unique challenges and opportunities compared to traditional privacypreserving computation: While demanding strict confidentiality and scalability around data volumes and data rates, industrial settings can benefit from publicly known stakeholders, which depend on their reputation to conduct business, easing the identification and sanctioning of misbehaviour.