Overview

This repository contains source code underlying our empirical analysis of traceability in the Monero blockchain.

Reproducing the deducibility analysis involves the following steps:

1. Export blockchain data from a fully synchronized Monero node
2. Import the data into a Neo4j database using the Neo4j batch importer
3. Create database indices and run the Sudoku algorithm
4. Querying the database

This repository does not contain the simulation of the different mixin-sampling procedures. It can be found at kvnl33/monero-simulations instead.

1. Monero Blockchain Data Export

• Install the Monero software, start the daemon and wait for the node to synchronize
• Install the requests python library: pip install requests
• Run the monero-to-csv.py script (i.e., python monero-to-csv.py)

2. Neo4j Import

• Download and/or install the Neo4j graph database v3.4.5
• You should now have a folder structure similar to this:
  • ./csv/: the Monero blockchain export (CSV files)
  • ./csv-headers/: the CSV header files
• Use the following import script to use the neo4j-import tool for a quick import (you'll need to modify the --into <DIRECTORY> directory to match the location of your Neo4j installation)

neo4j-import --into <DIRECTORY> --nodes:Block "csv-headers/blocks.csv,csv/blocks.csv" --relationships:PREV_BLOCK "csv-headers/blocks-rels.csv,csv/blocks-rels.csv" --nodes:Transaction "csv-headers/transactions.csv,csv/transactions.csv" --relationships:IN_BLOCK "csv-headers/tx-blocks.csv,csv/tx-blocks.csv" --nodes:Output "csv-headers/outputs.csv,csv/outputs.csv" --relationships:TX_OUTPUT "csv-headers/output-rels.csv,csv/output-rels.csv" --nodes:Input "csv-headers/inputs.csv,csv/inputs.csv" --relationships:TX_INPUT "csv-headers/input-rels.csv,csv/input-rels.csv" --relationships:REFERENCES "csv-headers/input-output-refs.csv,csv/input-output-refs.csv"

• Afterwards, you should see a message similar to the following:

IMPORT DONE in 5m 41s 347ms.
Imported:
  45169429 nodes
  84119631 relationships
  109326559 properties
Peak memory usage: 1.51 GB

3. Database Indexes and Mixin Sudoku

• Compile the Neo4j plugin in the neo4j-plugin folder: mvn package
• Put the resulting mixinsudoku.jar into the plugins folder of your Neo4j installation
• Download the APOC plugin and also put it into the plugins folder
• Allow unrestricted access by adding dbms.security.procedures.unrestricted=mixinsudoku.*,apoc.* to your Neo4j config
• Start the database
• Run the following cypher commands (through the web interface or the command line interface)
  • CALL mixinsudoku.schema(): creates all necessary indexes (you can check the status with :SCHEMA)
  • CALL mixinsudoku.coinbase(): adds labels for coinbase transactions
• Run the mixin sudoku in two steps (run the second query only after the first one has finished):
  1. CALL mixinsudoku.zeromixin(): labels all outputs spent by 0-mixin transactions
  2. CALL mixinsudoku.sudoku(): iteratively labels further deducable outputs
• Run CALL mixinsudoku.checkdb() as a sanity check at the end

4. Jupyter notebook

• Install the following Python modules:
  • pip install jupyter
  • pip install pandas
  • pip install seaborn
  • pip install py2neo
• Launch the notebook server: jupyter notebook &
• Open Monero Analysis.ipynb in the notebook interface