Analyzing various assumptions of the Stellar Consensus Protocol
Preliminary notes:
- The Stellar Consensus Protocol entirely rests on the assumption of "Quorum Intersection"
- Testing for Quorum Intersection is NP-complete; see: https://arxiv.org/pdf/1902.06493.pdf
- Why would you rest your entire consensus algorithm on an assumption you can't tractably test? ...
- Some analysis of the network dynamics of the Stellar Network has already been done; see: [TODO, can't find the paper now...]
TODO:
- Simulate networks with realistic parameters based on actual Stellar network activity.
- Compute how often Quorum Intersection actually holds. Note: this will need to be a lower bound due to intractability mentioned above.
References: