Performance Study for Improving Throughput in Hyperledger Fabric Blockchain Platform

Abstract

Hyperledger Fabric (HLF) is a blockchain platform that supports immediate finality of transactions and can be used in various application domains such as Supply chain, Health etc. Researchers have reported significant improvement in throughput, in HLF v1.0, based on experiments carried out with certain optimizations when Kafka is used as a consensus mechanism. In our paper, we supplement the above study with a few more guidelines that improve the throughput. The guidelines have been derived on top of HLF platform v2.2.3 with Raft and SmartBFT consensus mechanisms and with different Tx sizes. Our experimental study has shown 3 to 5 times improvement when compared to default configuration settings of HLF v2.2.3 in the case of Raft.