Header

Search

Agent-based modeling of Tendermint consensus protocol

Level: MA/MAP 

Responsible person: Dr. Matija Piškorec

Keywords: agent-based modeling, Tendermint, consensus protocols, Cosmos 

Description: 

Tendermint is a consensus protocol based on Byzantine Fault Tolerant (BFT) voting rules and Proof-of-Stake (PoS) as Sybil protection, used in Cosmos ecosystem of blockchains such as Cosmos Hub, Sei, Celestia, Osmosis, dYdX and many others. The goal of this project is to implement an agent-based model (ABM) of Tendermint consensus, calibrate it with empirical data and investigate its performance under various simulated contingency scenarios. Students can use an existing Python implementations of ABM models developed within the BDLT group as a starting point, including those for Stellar Consensus Protocol (SCP) [1,2,3], Proof-of-Work protocol [1,2] and Ethereum PoS [4,5]. Students should investigate ways of collecting necessary data for the calibration of Tendermint ABM simulator - peer-to-peer network between validators, inter-block time and number of transactions in each block (in order to approximate transaction finalization rate), optionally also mempool data on pending transactions not yet included in the blockchain. This data should be collected through publicly available APIs. The collected data should be then used for the calibration of the ABM model so that it reproduces empirical observables. Finally, students should develop and execute several contingency scenarios through which the performance of Tendermint consensus can be evaluated, such as high network latency, high transaction load, or presence of Byzantine validators. 

Reference:
[1] https://github.com/jaimedvw/stellar-simulator 

[2] Jaime de Vivero Woods, "Comparative Analysis of Stellar Consensus Protocol with Proof-of-Work Using Agent-Based Modelling," Master's thesis, University of Zurich, 2025. 

[3] Azizbek Asadov, Jaime de Vivero Woods, Jasur Fayzullaev, and Yukta Koralla, "Simulation and Analysis of the Stellar Consensus Protocol," Master project report, University of Zurich, 2025. 

[4] https://github.com/benckj/ethereum-consensus-abm 

[5] Kraner, Benjamin, et al. "Agent-based modelling of ethereum consensus." 2023 IEEE International Conference on Blockchain and Cryptocurrency (ICBC). IEEE, 2023.