Incentive-Compatible and Privacy-Preserving Data Analytics System enabled by Blockchain and Multiparty Computation

Blockchain, the technology at the foundation of Bitcoin and other cryptocurrencies, is being praised as a major disruptive innovation with the potential to transform most industries. Yet, considering its base architecture, it lacks some fundamental properties that are preventing a larger scale adoption. Among the others, being the blockchain a public and shared ledger, it doesn't offer privacy of data, precluding its usage to many applications that handle sensitive pieces of information. In this work, the blockchain is integrated with Secure Multi-Party Computation (MPC), a cryptographic tool that allows a number of parties to jointly compute a function over their inputs, keeping them private and guaranteeing that the output (if returned) is correct. Although the two might look apparently distant, they actually address the same set of problems, the ones where mutually mistrusting parties are involved without a trusted third party. Besides they cover different aspects of such problems, thus being considered complementary technologies. Blockchain (by means of Smart Contracts) shapes the MPC system to be incentive compatible, such that every computing party is incentivized to act honestly in order to maximize his profit, or alternatively penalized. It also provides a single and incorruptible source of truth of the entire system and, based only upon that, it enforces the contract conditions like an automated trusted third party. The resulting decentralized application (dApp) is implemented over a use-case belonging to the medical domain, in a scenario where patients earn cryptocurrency after having provided their sensitive data under an MPC scheme, while larger groups (i.e. research organizations) obtain aggregate data from arbitrary functions executed by the MPC system. Hopefully, the union between these two technologies could pave the way for a spectrum of new decentralized and incentive compatible applications, for many use-cases never explored before.