MPClan: Protocol Suite for Privacy-Conscious Computations

Abstract The growing volumes of data being collected and its analysis to provide better services are creating worries about digital privacy. To address privacy concerns give practical solutions, the literature has relied on secure multiparty computation techniques. However, recent research over rings mostly focused small-party honest-majority setting up four parties tolerating single corruption, noting efficiency concerns. In this work, we extend strategies support higher resiliency in an with online phase at centre stage. Our semi-honest protocol improves communication Damgård Nielsen (CRYPTO’07) without inflating overall communication. It also allows shutting down almost half phase, thereby saving 50% system’s operational costs. maliciously enjoys similar benefits requires only parties, except for one-time verification towards end, provides security fairness. showcase practicality designed protocols, benchmark popular applications such as deep neural networks, graph genome sequence matching, biometric matching using prototype implementations. addition improved communication, aid bringing 60–80% savings monetary cost prior work.