Accelerating Hyperledger Fabric: A Dual-Cache Architecture for Efficient Block Submission in IoT Systems

Blockchain technology, characterized by its distributed, tamper-proof, and encrypted nature, offers robust solutions for privacy and data security in the Internet of Things Systems (IoTSys). However, the high concurrency requirements of IoTSys often overwhelm the throughput capacity of existing blockchain 3.0 systems like Hyperledger Fabric, which lags behind traditional centralized databases. Through a granular analysis of the HLF architecture, we identified that synchronous disk I/O during the critical path of block submission significantly increases latency and throttles throughput. To address this, we propose a high-performance framework comprising two core strategies: a Cache-Based Block Submission Process Optimization Strategy and a World State Semi-Synchronous Write Strategy. By introducing an intermediary caching layer and decoupling state persistence from the transaction execution path, our design mitigates redundant disk I/O. Experimental results under high client load scenarios demonstrate that our architecture improves throughput by approximately 10% and reduces transaction latency by 25% compared to native HLF, without compromising data consistency or system availability.