Recent Results in Scalable Multi-Party Computation

Secure multi-party computation (MPC) allows multiple parties to compute a known function over inputs held by each party, without any party having to reveal its private input. Unfortunately, traditional MPC algorithms do not scale well to large numbers of parties. In this paper, we describe several recent MPC algorithms that are designed to handle large networks. All of these algorithms rely on recent techniques from the Byzantine agreement literature on forming and using quorums. Informally, a quorum is a small set of parties, most of which are trustworthy. We describe the advantages and disadvantages of these scalable algorithms, and we propose new ideas for improving practicality of current techniques. Finally, we conduct simulations to measure bandwidth cost for several current MPC algorithms.