A Blockchain‐Based Batched Data Sharing Protocol for Efficient AIoT Communications

ABSTRACT

In AIoT‐driven applications such as precision healthcare, multiple devices communicate through continuous sharing of fragmented and sovereignty‐constrained data. This poses critical challenges in enabling secure data sharing among multiple trustless stakeholders under fine‐grained access control policies. Although blockchain provides a user‐sovereign and trust‐enhancing framework, existing work remains inefficient in healthcare scenarios with frequent recipient and data changes due to repeated operations. It requires re‐authorizing every authorized segment for each recipient and relies on non‐batchable mechanisms for transaction atomicity, resulting in additional computational cost and on‐chain transaction overhead. To address these limitations, we propose the Dynamic Authorization with Batched Transaction Atomicity (DABTA) protocol that achieves dynamic authorization with fixed overhead and batched data sharing transaction atomicity. DABTA introduces a constant‐size combination‐bound artifact to decouple authorized recipients from authorized content, eliminating repeated encryption when either the shared data or the recipient set changes. Furthermore, DABTA aggregates multiple authorization transactions under a shared commit checkpoint and performs batched decryption to recover transaction payloads, thereby amortizing the overhead of ensuring transaction atomicity across concurrent authorizations. We prove the correctness and security of DABTA. A prototype built on Hyperledger Fabric achieves 97.4 authorizations per second, delivering approximately 68.2% higher throughput than a non‐batched baseline.