Paper 2023/511 Original: eprint.iacr.org/2023/511 ↗

$\text{MP}\ell\circ \mathrm{C}$: Privacy-Preserving IP Verification Using Logic Locking and Secure Multiparty Computation

Abstract

The global supply chain involves multiple independent entities, and potential adversaries can exploit different attack vectors to steal proprietary designs and information. As a result, intellectual property (IP) owners and consumers have reasons to keep their designs private. Without a trusted third party, this mutual mistrust can lead to a deadlock where IP owners are unwilling to disclose their IP core before a financial agreement is reached, while consumers need assurance that the proprietary design will meet their integration needs without compromising the confidentiality of their test vectors. To address this challenge, we introduce an efficient framework called MPloC that resolves this deadlock by allowing owners and consumers to jointly evaluate the target design with consumer-supplied test vectors while preserving the privacy of both the IP core and the inputs. MPloC is the first work that combines secure multiparty computation (MPC) and logic-locking techniques to accomplish these goals. Our approach supports both semi-honest and malicious security models to allow users to balance stronger security guarantees with performance. We compare our approach to existing state-of-the-art works that utilize homomorphic encryption across several benchmarks and report runtime improvements of more than two orders of magnitude.