Fourteen years after its inception, the Bitcoin market capitalization exceeds $700 trillion. As Bitcoin blockchain continues gaining tremendous interest, it is vital evaluating Bitcoin protocol performance. In this context, the BlockSim simulation framework is among the current state-of-the-art tools. Despite that, the BlockSim model of the Bitcoin protocol shows two main limitations: i) at the consensus layer, the model does not account for Segregated Witness (SegWit) upgrade, which sensibly improves the Bitcoin throughput; and ii) at the network layer, the simulated block propagation process is oversimplified. Those limitations affect several blockchain performance metrics, like throughput, orphan block rate and mining reward per block. In this work we improve the existing BlockSim model of the Bitcoin protocol. Namely, we introduce SegWit support, and we update the simulated block propagation process. Block propagation delay is now estimated via linear regression on a per-simulated-block basis. We also extensively validate the proposed model within the light simulation technique of BlockSim. The results show that the model correctly simulates the current Bitcoin blockchain. As to the full simulation technique of BlockSim, it actually turns out to be so inefficient to be unpractical. We experimentally prove the performance limitations of the full technique, thereby confirming the original claims regarding its inefficiency. Yet, we provide suggestions to improve execution time and memory footprint, corroborated by profiling results.

"Keywords: {Blockchain, BlockSim, Bitcoin, Segregated Witness, Simulation, Validation}"
"File: https://ieeexplore.ieee.org/document/9881797"