Paper 2018/629

Context Hiding Multi-Key Linearly Homomorphic Authenticators

Lucas Schabhüser, Denis Butin, and Johannes Buchmann

Abstract

Demanding computations are increasingly outsourced to cloud platforms. For such outsourced computations, the efficient verifiability of results is a crucial requirement. When sensitive data is involved, the verification of a computation should preserve the privacy of the input values: it should be context hiding. Context hiding verifiability is enabled by existing homomorphic authenticator schemes. However, until now, no context hiding homomorphic authenticator scheme supports multiple independent clients, e.g. multiple keys. Multi-key support is necessary for datasets involving input authenticated by different clients, e.g. multiple hospitals in e-health scenarios. In this paper, we propose the first perfectly context hiding, publicly verifiable multi-key homomorphic authenticator scheme supporting linear functions. Our scheme is provably unforgeable in the standard model, and succinct. Verification time depends only linearly on the number of clients, in an amortized sense.