TinyOLE

Efficient Information Theoretic Multi-party Computation from Oblivious Linear Evaluation

“…4. Lastly, unlike the TinyOLE scheme [7], our scheme is scalable, in that it extends to the multi-party case with n participants. In fact, computing n shares (one for each participant) to a single multiplication requires only n − 1 OLEs, one for each individual participant to compute his share.…”

“…It suffices to assume that participants gain the shares of the triples via an external party known as an initialiser, who (after computing and distributing the shares of the triples) does not take part in the actual MPC protocol. In the next section we review the OLE based two-party protocol given by Döttling et al [7].…”

“…TinyOLE Recently Döttling et al [7] proposed a two-party protocol (n = 2) in which the two participants, P 1 and P 2 use OLE to compute shares to a multiplication. Specifically, they use OLE to compute multiplication triples in an offline phase.…”