CLOAK: Enabling Confidential Smart Contract With Multi-Party Transactions

In recent years, as blockchain adoption has been expanding across a wide range of domains, e.g., supply chain finance, digital asset, etc., the confidentiality of smart contracts has now become a fundamental demand for practical applications. However, while new privacy protection techniques are emerging, how existing ones can best fit development settings are understudied. State-of-the-art solutions lack architectural support - in terms of programming interfaces - thus are hardly able to reach general developers. This paper proposes Cloak, a pluggable and configurable framework for developing and deploying confidential smart contracts. The key capability of Cloak is to allow developers to implement and deploy practical solutions to \emph{multi-party transaction} (MPT) problems, i.e., to transact with secret inputs and states owned by different parties, by simply \emph{specifying} it. To this end, Cloak allows users to specify privacy invariants in a declarative way, automatically generate runtime with verifiably enforced privacy and deploy it to the existing platforms with TEE-Blockchain architecture, enabling the MPT. Additionally, we identify the pitfalls in achieving MPT, and provide the treat, i.e., non-deterministic negotiation and fair publication of MPT results. In our evaluation on both examples and real-world applications, developers manage to deploy business services on blockchain in a concise manner by only developing Cloak smart contracts, whose size is less than 30% of the deployed ones, and the gas cost of deployed MPTs reduced by 19%. We believe the insights learned from Cloak will pave the way for general-purpose multi-party privacy-preserved computation achieved by harmonizing TEE and blockchain.