Performance Modeling of Hyperledger Fabric 2.0

Hyperledger Fabric has become one of the most widely used consortium blockchain frameworks with the ability to execute custom smart contracts. Performance modeling and network evaluation are necessary for performance estimation and optimization of the Fabric blockchain platform. The compatibility and effectiveness of existing performance modeling methods must be improved. For this reason, we proposed a compatible performance modeling method using queuing theory for Fabric considering the limited transaction pool. Taking the 2.0 version of Fabric as a case, we have established the model for the transaction process in the Fabric network. By analyzing the two-dimensional continuous-time Markov process of this model, we solved the system stationary equation and obtained the analytical expressions of performance indicators such as the system throughput, the system steady-state queue length, and the system's average response time. We collected the required parameter values through the official test suite. An extensive analysis and simulation was performed to verify the accuracy and the effectiveness of the model and formula. We believe that this method can be extended to a wide range of scenarios in other blockchain systems.