Performance Evaluation of M/D/1 Queuing Model on Hyperledger Fabric

Abstract

Hyperledger Fabric is a private blockchain intended to be confidentiality, security, scalability, transparency, and access control service to the enterprise. The impact of the configuration factors influences the performance of Hyperledger Fabric because of a modular architecture. It does not provide to find the performance metrics of each processing phase likes endorsing peer, ordering service, and committing peer through the experiments of the formal Software Development Kit (SDK). In this paper, the performance metric of ordering service is presented in Hyperledger Fabric by using M/D/1 queuing model. The utilization is evaluated based on the random transaction arrival rate to a single orderer node and the deterministic service rate to determine the addition of a new orderer node. The number of orderer nodes could be decided based on the model result to improve the real system performance.