We describe Chained Shamir Batching, a method for ensuring the integrity of audit logs without relying on any single trusted party. The method combines Merkle trees for efficient log aggregation, Shamir's Secret Sharing for distributed key custody, and a chaining mechanism that cryptographically links successive batches. The resulting chain is self-verifying: tampering with any batch requires tampering with all subsequent batches, while independent backup keys provide a second verification path. Unlike blockchain-based solutions, the method requires no consensus mechanism and introduces negligible computational overhead. Unlike centralized approaches, no single entity - neither the log-producing organization, the storage provider, nor the auditor - can unilaterally alter, delete, or access log contents. The approach is designed to meet the provable integrity requirements of regulatory frameworks such as NIS2.
Vladislav Urbánek (2026) studied this question.