Performance Modeling of Hyperledger Fabric (Permissioned Blockchain Network)

Abstract

Hyperledger Fabric (HLF) is an open-source implementation of a distributed ledger platform for running smart contracts in a modular architecture. In this paper, we present a performance model of Hyperledger Fabric v1.0+ using Stochastic Reward Nets (SRN). From our detailed model, we can compute the throughput, utilization and mean queue length at each peer and critical processing stages within a peer. To validate our model, we setup an HLF network in our lab and run workload using Hyperledger Caliper. From our analysis results, we find that time to complete the endorsement process is significantly affected by the number of peers and policies such as AND (). The performance bottleneck of the ordering service and ledger write can be mitigated using a larger block size, albeit with an increase in latency. For the committing peer, the transaction validation check (using Validation System Chaincode (VSCC)) is a time-consuming step, but its performance impact can be easily mitigated since it can be parallelized. However, its performance is critical, since it absorbs the shock of bursty block arrivals. We also analyze various what-if scenarios, such as peers processing transactions in a pipeline, and multiple endorsers per organization.