Confidential computing

The possibility of exploiting a large number of distributed resources to analyse heterogeneous data, especially for medical or biological applications, is hampered by the fact that often such data contain sensitive information or is protected by policies and usage agreements that do not allow the data to be shared. However, the recent pandemic and the many unsolved health challenges (HIV, neurological and mental conditions, overstressed healthcare systems, to mention a few) require a different approach to data collection, sharing and analysis. Traditional approaches like (pseudo-) anonymisation require compromises between privacy preservation and information content preservation.
Novel techniques to encrypt the data (such as homomorphic encryption) or move the data across provably secure connections might turn the tide. Projects like openQKD, part of the European Quantum Flagship strategy, and others are investigating the feasibility of large-scale QKD networks. CERN is currently involved in the development of an end-to-end QKD-based secure data analytics demonstrator, Quantumacy, that proposes to advance the application of cryptographic techniques, federated learning and QKD to distributed data processing.
Discussions to set up joint research in this area are now taking place with infrastructure providers, such as GEANT in Europe, and domain experts in the industry (IDQ) and research (PSNC in Poznań).