Level: Master Level
Responsible Person: Parminder Kaur Makode (parminderkaur.makode@uzh.ch)
Type of work: Agent-based Modelling (Python)

The Stellar Consensus Protocol (SCP) [1] is a consensus algorithm used by the Stellar network, which is a decentralized, open source blockchain platform. SCP is designed to reach agreement among distributed nodes in the network on the validity and ordering of transactions.
This thesis focuses on modeling the Stellar Consensus Protocol (SCP) and devising metrics to assess the consensus quality. Building upon the research group’s previous work, we aim to develop an extended model that captures the essence of SCP. Our expertise lies in analyzing and modeling consensus protocols, particularly in the context of information diffusion within overlay P2P networks. To illustrate, our previous studies [2] employed emulation to examine efficiency issues in large-scale Ethereum networks, while [4] introduced a minimalist stochastic model applicable to consensus in blockchain-based systems, encompassing both Proof-of-Work (PoW) and Proof-of-Stake (PoS) variants, akin to epidemic models.
The primary objective is to comprehend the scalability and capabilities of the existing Stellar Consensus Protocol by constructing a tailored model that allows investigation of its current properties and potential modifications required to achieve specific objectives. For instance, seeking to understand how the protocol behaves under network congestion and exploring potential remedies for such scenarios. Additionally, the aim is to analyze the structural characteristics of the Stellar P2P network and how they evolve across various modeling scenarios.
To achieve this, a model of SCP utilizing the Federated Byzantine Agreement (FBA) for consensus will be created [3] [1]. Since the transaction order is crucial to SCP’s functionality, the model will incorporate the incoming transaction rate as a significant feature. The work will also investigate how the transaction order influences the concept of quorum slices. Moreover, the model will encompass P2P network topology and incorporate difference consensus mechanisms such as balloting, nominating, and federated voting.

Reference:
[1] D. Mazieres, "Stellar Consensus Protocol," [Online]. Available: https://stellar.org/papers/stellar-consensus-protocol.
[2] C. J. T. M. V. S. V. D. G. M. E. M. Maximiliano Geier, "Using Network Emulation to Study Blockchain Distributed Systems: The Ethereum Case," in 27th Euromicro International Conference on Parallel, Distributed and Network-based Processing (PDP), 13-15 Feb. 2019.
[3]G.L.D.M.G.H.N.B.E.G.J.J.R.M.a.J.M.MartaLokhava,"FastandSecure Global Payments with Stellar," SOSP ACM, pp. 80-96, 2019.
[4]P. T. F. I. Claudio J. Tessone, "Stochastic Modelling of Blockchain Consensus," Distributed, Parallel, and Cluster Computing, 2021.