SVD-BTR: a lightweight storage–verification decoupling architecture with a blockchain-anchored trusted root for trusted data provenance

Abstract

Trusted data provenance is increasingly critical to protect against sensitive data leakage in defense, critical infrastructure, and high-tech environments. Ensuring trusted data provenance at scale is challenging because existing on-chain/off-chain designs reduce on-chain storage but still incur chain-length dependent verification costs due to repeated on-chain queries. We propose SVD-BTR, a lightweight architecture that enables full-chain integrity verification with a single on-chain check. SVD-BTR uses storage–verification decoupling (SVD) and treats the tail record in a backtracking query as a trusted verification root (TVR) whose compact commitment is anchored on blockchain. This shifts the blockchain from a record-level hash database to a compact trust anchor: after local provenance reconstruction, SVD-BTR requires only a single blockchain-side TVR check—achieving O(1) on-chain interaction per query—while local computation remains path-length dependent. Experiments show that SVD-BTR improves high-concurrency write throughput by 14.4%, increases execution efficiency by 45.6% in long-chain scenarios (200 records), and reduces on-chain storage overhead by 75.4%.