Blockchain-Based Secure and Composable Model for Semi-Honest Privacy-Preserving Computation

Abstract

Privacy-preserving computation enables multiple parties to jointly compute a function while keeping their inputs private. Protocols designed for the semi-honest model achieve high efficiency by assuming participants will correctly follow the protocol’s cryptographic steps. However, this security assumption is confined to the protocol’s internal execution, creating a crucial accountability gap. It offers no inherent method to prove that the inputs and function used in the computation actually align with what was externally agreed upon. In this paper, we introduce a novel framework that enhances privacy-preserving computation with public verifiability and accountability, while maintaining composability. Our framework leverages a blockchain as an immutable trust anchor to register cryptographic commitments of both participant inputs and the function’s specification. We then employ a zero-knowledge proof protocol to verify that the privacy-preserving computation is performed correctly using the committed data and function logic. The security of our model is formally proven to guarantee both input privacy and computational integrity, while our performance evaluation shows its practical scalability.