An Overlapping Sharding Optimization Mechanism for Blockchain in IIoT

Tianyi Xu, Xinjie Yang, Xiaobo Zhou, Ziqi Gan, Lin Gan, Tie Qiu

Abstract: Sharding is widely concerned to address the scalability limitations of blockchain in Industrial Internet of Things (IIoT). However, the scalability and throughput improved by using sharding are degraded by the validation of crosss-hard transactions. Recent studies have proposed overlapping sharding architecture that converts cross-shard transactions into overlap-shard transactions, which processes cross-shard transaction validation within the over-lapping part between shards, thereby greatly reduces the processing overhead. In the over-lapping sharded blockchain, the performance of nodes of the overlapping parts is crucial to the security and throughput of the system. Because devices and servers in IIoT suffer from weak computing and storage capacity and are vulnerable to attacks. To allocate nodes to different overlapping shards to enhance cross-shard transaction validation and ensure security regarding the heterogeneity of devices remains as a critical problem for blockchain-based IIoT systems. To address that, we balance multiple factors affecting the throughput and security of sharded blockchain-based IIoT system, and abstract the overlapping shards formation problem as a multi-objective optimization problem. In addition, we propose a novel Overlapping Sharding Optimization Mechanism based on genetic algorithms (OSOM) to solve it. Experimental results indicate that OSOM enhances both the security and performance of overlapping sharding systems in blockchain systems for IIoT.