MPC-ABC: Blockchain-Based Network Communication for Efficiently Secure Multiparty Computation

Secure Multiparty Computation (MPC) offers privacy-preserving computation that could be critical in many health and finance applications. Specifically, two or more parties jointly compute a function on private inputs by following a protocol executed in rounds. The MPC network typically consists of direct peer-to-peer (P2P) connections among parties. However, this significantly increases the computation time as parties need to wait for messages from each other, thus making network communication a bottleneck. Most recent works tried to address the communication efficiency by focusing on optimizing the MPC protocol rather than the underlying network topologies and protocols. In this paper, we propose the MPC over Algorand Blockchain (MPC-ABC) protocol that packs messages into Algorand transactions and utilizes its fast gossip protocol to transmit them efficiently among MPC parties. Our approach, therefore, reduces the delay and complexity associated with the fully connected P2P network while assuring the integrity of broadcasted data. We implemented MPC-ABC and utilized it to outsource the SPDZ (SPDZ-pronounced "Speedz"-is the nickname of the MPC protocol of Damgard et al. in (European Symposium on Research in Computer Security, pp 1-18, 2013)) protocol across multiple Cloud Service Providers (CSP). Experimental results show that our approach outperforms the commonly adopted approaches over the P2P TCP/IP network in terms of the average delay and network complexity.