A practical model for collaborative databases: Securely mixing, searching and computing

Abstract

We introduce the notion of a Functionally Encrypted Datastore which collects data anonymously from multiple data-owners, stores it encrypted on an untrusted server, and allows untrusted clients to make select-and-compute queries on the collected data. Little coordination and no communication is required among the data-owners or the clients. Our notion is general enough to capture many real world scenarios that require controlled computation on encrypted data, such as is required for contact tracing in the wake of a pandemic. Our leakage and performance profile is similar to that of conventional searchable encryption systems, while the functionality we offer is significantly richer. In more detail, the client specifies a query as a pair (Q, f) where Q is a filtering predicate which selects some subset of the dataset and f is a function on some computable values associated with the selected data. We provide efficient protocols for various functionalities of practical relevance. We demonstrate the utility, efficiency and scalability of our protocols via extensive experimentation. In particular, we evaluate the efficiency of our protocols in computations relevant to the Genome Wide Association Studies such as Minor Allele Frequency (MAF), Chi-square analysis and Hamming Distance.