Bachelor/Master Thesis
Information-Theoretic Partially Synchronous BFT
Byzantine Fault Tolerant (BFT) protocols are a fundamental building block for distributed systems operating in adversarial environments. Classical protocols typically rely on cryptographic assumptions to ensure safety and liveness. However, an alternative line of work studies information-theoretic BFT, where correctness is guaranteed without relying on computational hardness assumptions.
Several protocols have been proposed for the partially synchronous setting [1, 2, 3, 4, 5]. These algorithms strive to obtain the best possible guarantees in the following measurements, while not relying on cryptographic tools like digital signatures.
- Storage requirements
- Communication complexity
- Responsiveness
- Good-case latency
Until recently, the only protocol that provided optimal guarantees for all measurements except latency was TetraBFT, which has a good-case latency of 5 rounds. More recently, Forget-It [1] was able to provide optimal guarantees for all properties, including an optimal good case latency of 3 rounds.
In this project, we will study the emerging literature on information-theoretic partially synchronous BFT protocols. The objective of the thesis is to build a solid understanding of these protocols, analyze their guarantees and limitations, and explore the design principles that enable efficiency without cryptographic assumptions.
References
[1] Forget-It: Efficient Information-Theoretic Partially Synchronous BFT [2] TetraBFT: Reducing Latency of Unauthenticated, Responsive BFT Consensus [3] Information Theoretic HotStuff [4] Stellar Consensus Protocol [5] Practical Byzantine Fault Tolerance