Semi-honest adversaries

Abstract

In this chapter, we present Yao’s protocol for secure two-party computation in the presence of semi-honest adversaries. The protocol has a constant number of rounds, and works by having the parties evaluate an “encrypted” or “garbled” circuit such that they learn nothing from the evaluation but the output itself. In particular, all intermediate values in the circuit evaluation (which can reveal more information than is allowed) remain hidden from both parties. We present the protocol for the case of a deterministic, non-reactive, single-output functionality. As we have shown in Section 2.5, this suffices for obtaining the secure computation of any probabilistic, reactive two-party functionality at approximately the same cost.