The goal of this thesis is to study security properties of the JoinMarket CoinJoin protocol [1] for Bitcoin privacy mixing and its practical implementation. In the theoretical part, the current protocol concepts and implemented features will be described based on the (limited) documentation and available open-source code. Known methods for anonymity estimation and applicable anonymity-decreasing attacks will be researched and summarized, including the so-called CoinJoin Sudoku attack [2]. In the practical part, a combination of real data analysis (JoinMarket orderbook and mainnet transactions) and a working Dockerized setup with modified JoinMarket clients [3] for large-scale emulation will be used. An extensive emulations with known ground truth will be executed and analyzed to evaluate expected anonymity gain as well as the impact of the attacks considered. A discussion of results, together with recommendations for attack mitigations, will be provided.

The following practical outputs shall be covered:

• Adaptation of JoinMarket clients and coordination for the emulation of a Dockerized setup.
• At least three large-scale scenarios will be emulated with relevant data collected and analyzed. 
• Evaluation of CoinJoin Sudoku attack impact on relevant privacy metrics using emulation on the regtest network.
• Capture and analysis of real JoinMarket orderbook data.
• Analysis of real JoinMarket transactions extracted from the Bitcoin mainnet blockchain with a discussion of the implications of findings from the emulation setup. 

Literature:

[1] JoinMarket design, https://github.com/JoinMarket-Org/JoinMarket-Docs/blob/master/High-level-design.md

[2] K. Atlas, CoinJoin Sudoku, https://github.com/ARLM-Attic/coinjoin-sudoku

[3] JoinMarket client software, https://github.com/JoinMarket-Org/joinmarket-clientserver