Paper 2024/1184 Original: eprint.iacr.org/2024/1184 ↗

Sanitizable and Accountable Endorsement for Dynamic Transactions in Fabric

Abstract

Hyperledger Fabric, an open-source, enterprise-grade consortium platform, employs an endorsement policy wherein a set of endorsers signs transaction proposals from clients to confirm their authenticity. The signatures from endorsers constitute the core component of endorsement. However, when dealing with dynamic transactions with high timeliness and frequent updates (e.g., stock trading, real-time ad delivery, news reporting, etc.), the current endorsement process somewhat slows down the transaction execution. Meanwhile, handling these continuously updated transactions consumes significant resources from endorsers, thereby constraining overall application efficiency. To address these issues, this paper devises a novel sanitizable and accountable endorsement scheme by proposing a sanitizable multi-signature (SMS) as the theoretical tool. Specifically, we introduce the novel concept of sanitizable multi-signature and detail its instantiation. SMS combines the advantages of multi-signature and sanitizable signature, maintaining the compactness of the signature while allowing the sanitizer to adjust the initial endorsement result to fit the updated transaction content without interacting with the endorsers, so that both the authenticity and timeliness of transactions can be ensured. Additionally, SMS incorporates an innovative accountability mechanism to trace instances of improper data updates, thereby enhancing the security and reliability of the endorsement process. We demonstrate the security of the proposed scheme through rigorous security analysis. Performance evaluations show that SMS can significantly reduce verification overhead and transaction size compared to the default ECDSA scheme in Fabric. Specifically, when verifying multiple endorsers' endorsements, our scheme exhibits a storage space reduction by approximately 30%-40% and a verification time reduction ranging from 9.2% to nearly 26.3%.