LRC-DPoR: Efficient Data Integrity Auditing with Local Repair and Dynamic Updates

Proof of Retrievability (PoR) enables verifiable outsourced storage with the guarantee that an entire file can be recovered from a possibly untrusted server. However, most of the dynamic PoR (DPoR) schemes that support data modification, insertion, and deletion incur high update complexity and excessive repair bandwidth, largely due to their reliance on global erasure codes and position-dependent authentication. In this work, we present LRC-DPoR, a dynamic PoR construction that integrates Local Reconstruction Codes (LRC) into the PoR framework to achieve provable retrievability with asymptotically improved repair and update efficiency. Our scheme partitions encoded data into locally repairable groups, enabling exact repair of single-block failures by accessing only a bounded number of blocks, while retaining global recoverability under multiple failures. To support full data dynamics without re-encoding or authentication tag shifting, we introduce an out-of-place, append-only update model together with an authenticated array-based index structure, which separates logical ordering from physical storage and supports exact modifications, insertions, and deletions. We prove that LRC-DPoR is unforgeable under standard cryptographic assumptions, and imply original data is recoverable via the underlying LRC-based encoding. Analytical evaluation demonstrates that our scheme achieves strictly low asymptotic repair bandwidth and update overhead, while preserving strong retrievability guarantees.