The smart contract model of asynchronous, robust MPC

Abstract

Over the past few years, massive cyberattacks have dominated the public imagination. Most
\nAmericans have been directly affected by data breaches and an overwhelming majority believe
\nthat they are no longer in control of their data. At the same time, there are important applications in
\nwhich the aggregation of private data is unavoidable. Not surprisingly, there has been considerable
\ninterest in developing protocols for secure multi-party computation (MPC), i.e., N parties providing
\nprivate inputs to jointly compute some function f. Recent developments in MPC have led to leaps
\nin efficiency and MPC is quickly becoming a practical approach for privacy-centric distributed
\napplications. As MPC applications become deployed, guaranteeing the security of these protocols,
\neven when interacting with other applications, is essential. The standard approach to showing
\nsecurity of MPC protocols under arbitrary composition is through the universal composability (UC)
\nframework. There has been much prior work on this topic; however, most previous work either
\nmakes synchronicity assumptions or does not guarantee output delivery in the presence of even
\none fault. For practical MPC applications, though, dealing with asynchronicity and robustness is
\nessential. In this thesis, we describe two new keywords, eventually and leak, that aid in defining
\nprotocols in the asynchronous world and define wrappers that implement these within the UC
\nframework. We then use this novel wrapper to give a UC-realization of a reliable broadcast primitive, by means of Bracha’s classic protocol. Finally, we define and prove a realization of what we call the
\nsmart contract model of MPC, which serves as a UC-idealization of asynchronous, robust MPC.