A Toy MPC Protocol

Wenxuan Yu, Minghui Xu, Bing Wu +2 more

MPC-as-a-Service (MPCaaS) systems enable clients to outsource privacy-preserving computations to distributed servers, offering flexibility by adapting and configuring MPC protocols to meet diverse security requirements. However, traditional MPC protocols rely on a fixed set of servers for the entire computation process, limiting scalability. Dynamic MPC (DMPC) addresses this limitation by permitting participants to join or leave during the computation. Nevertheless, existing DMPC protocols assume synchronous networks, which can lead to failures under unbounded network delays. In this paper, we present AD-MPC, the first asynchronous dynamic MPC protocol. Our protocol ensures guaranteed output delivery under optimal resilience ((n = 3t + 1)). To achieve this, we introduce two critical components: an asynchronous dynamic preprocessing protocol that facilitates the on-demand generation of Beaver triples for secure multiplication, and an asynchronous transfer protocol that maintains consistency during party hand-offs. These components collectively ensure computation correctness and transfer consistency across participants. We implement AD-MPC and evaluate its performance across up to 20 geographically distributed nodes. Experimental results demonstrate that the protocol not only offers strong security guarantees in dynamic and asynchronous network environments but also achieves performance comparable to state-of-the-art DMPC protocols.

Nan Cheng

Marcel Keller, Dragos Rotaru, Nigel Paul Smart +1 more

Yunqi Yang, Rui Zhang

Secure multiparty computation (MPC) requires the messages transmitted in secure channels which can provide encryption and authorization to the messages. To implement a secure channel for MPC protocols, researchers have tried some communication protocols, such as TLS and Noise. However, these methods have some limitations. These protocols need a trusted certification authority to provide identity authorization which is difficult for an MPC protocol, and how participants manage the key of each party and how to use the key to establish communication is also a problem. A Signal protocol is an end-to-end encryption communication protocol, which is known as the most secure communication protocol in the world. Based on the Signal protocol, we implemented a signal-based secure multiparty computation protocol, which can run the MPC protocol and transmit messages through a signal-based secure channel. Compared with previous research on the MPC protocol over Signal secure channels, the new MPC client adds group communication of the Signal protocol to transmit messages, which significantly improves the communication efficiency of broadcast messages of MPC protocols. To test the communication efficiency of the new MPC client, we implemented a concrete BLS threshold signature protocol on the new client, comparing the elapsed time of key generation and signing on the new client to that on the client only using Signal end-to-end communication. According to our experiment result, we found that the new client run at least 37.48% faster than the old client on the BLS threshold signature whose number of parties ranges from 3 to 5, if the parties have sent Signal group messages to each other. The more parties in the MPC protocol, the higher the proportion of broadcast messages and the more obvious the performance improvement of the new client. Our work improves the performance of MPC secure channels based on the Signal protocol, especially for complex MPC protocols with many participants.

Yibiao Lu, Bingsheng Zhang, Kui Ren

Most existing MPC protocols consider the homogeneous setting, where all the MPC players are assumed to have identical communication and computation resources. In practice, the player with the least resources often becomes the bottleneck of the entire MPC protocol execution. In this work, we initiate the study of so-calledload-balanced MPCin heterogeneous computing. A load-balanced MPC protocol can adjust the workload of each player accordingly to maximize the overall resource utilization. In particular, we propose new notions calledcomposite circuitandcomposite garbling scheme, and construct two efficient server-aided protocols with malicious security and semi-honest security, respectively. Our maliciously secure protocol is over$400\times $faster than the authenticated garbling protocol (CCS ’17) and up to$4.3\times $faster than the state-of-the-art server-aided MPC protocol of Lu et al. (TDSC ’23); our semi-honest protocol is up to$173\times $faster than the optimized BMR protocol (CCS ’16) and is up to$3.8\times $faster than the protocol of Lu et al.

Bar Alon, Eran Omri, Anat Paskin-Cherniavsky