Early-stage Conflict Detection in HLF-based Delay-critical IoT Networks

Hyperledger Fabric (HLF), a permissioned blockchain network recognized for high throughput and low transaction latency, has proven instrumental in augmenting security and privacy. However, with the surge in Internet of Things (IoT) devices, the volume of transactions relayed to the HLF blockchain has grown, leading to the problem of conflicting transactions. This problem occurs when two transactions attempt simultaneous read and write operations on the same ledger key. Given the time-sensitive nature of certain delay-critical IoT networks, these conflicts, resolved during the validation stage, can adversely affect key network performance metrics, including throughput and transaction latency. Current solutions, while addressing these conflicts, often fail to detect them at an early stage, resulting in performance degradation. To address this problem, we introduce a novel Early-stage Conflict Detection (ECD) mechanism. Designed specifically for HLF-based delay-critical IoT networks, the ECD mechanism employs a local cache, namely Sync.Map, to store and compare incoming transactions, enabling earlier conflict detection. Our results show that the ECD mechanism significantly improves valid throughput, transaction latency, and conflict detection time compared to existing approaches.