Hyperledger Fabric‐Based Biometric Identity Management and Authentication: A Performance‐Centric Approach to Block Size Optimization

ABSTRACT Blockchain technology has emerged as a transformative solution to the limitations of centralized data management, offering transparency and decentralization. However, optimizing block size remains a critical challenge that directly influences performance metrics such as latency and throughput. This study develops an analytical model for determining the optimal block size in Hyperledger Fabric (HF) by formulating its dependence on endorsement policy, transaction volume, and read/write set dimensions. The model quantitatively expresses block composition and predicts performance variations under different network configurations. To validate the analytical findings, the model is applied to a Blockchain‐based Biometric Identity Management and Authentication System (BIMAS) implemented on HF. Hyperledger Caliper is utilized for the experimental evaluations on the benchmark FVC2004 (DB1, DB2, DB3, DB4) fingerprint databases. The effects of block size are evaluated on the key performance metrics, including latency, throughput, CPU, and memory consumption. The empirical results confirm the analytical predictions, revealing an optimal block‐size range that achieves minimal latency at 50 Transactions per second (TPS); at 250 TPS, it peaks for block size 20 and decreases as block size increases. Throughput scales near‐linearly with transaction arrival rate up to 150 TPS and then saturates, flattening by 200 TPS. The findings provide practical design guidance for developers deploying performance‐sensitive Blockchain applications, demonstrating how block‐size optimization can significantly enhance the scalability and responsiveness of HF in real‐world identity authentication systems.