Performance Evaluation of Hyperledger Fabric with Malicious Behavior

Abstract

Hyperledger Fabric is a widely-used permissioned blockchain platform for enterprise consortium applications. It adopts Practical Byzantine Fault Tolerance (PBFT) algorithm as the consensus protocol in its version 0.6. Faulty replicas could intentionally delay messages, be not responsive and send inconsistent messages to different replicas. Faulty clients and replicas could also launch denial-of-service attack to make resources unavailable. The malicious behavior significantly undermines the system. However, the existing performance evaluation for Fabric is accomplished in a fault-free environment without malicious behaviors. In this paper, we analyze the impact of malicious behavior, design malicious behavior patterns and test the blockchain performance on Hyperledger Fabric.