TetraBFT: Reducing Latency of Unauthenticated, Responsive BFT Consensus

Qianyu Yu,Giuliano Losa,Xuechao Wang

2024-05-04

Abstract:This paper presents TetraBFT, a novel unauthenticated Byzantine fault tolerant protocol for solving consensus in partial synchrony, eliminating the need for public key cryptography and ensuring resilience against computationally unbounded adversaries. TetraBFT has several compelling features: it necessitates only constant local storage, has optimal communication complexity, satisfies optimistic responsiveness -- allowing the protocol to operate at actual network speeds under ideal conditions -- and can achieve consensus in just 5 message delays, which outperforms all known unauthenticated protocols achieving the other properties listed. We validate the correctness of TetraBFT through rigorous security analysis and formal verification. Furthermore, we extend TetraBFT into a multi-shot, chained consensus protocol, making a pioneering effort in applying pipelining techniques to unauthenticated protocols. This positions TetraBFT as a practical and deployable solution for blockchain systems aiming for high efficiency.

Cryptography and Security

What problem does this paper attempt to address?

### What problems does this paper attempt to solve?

This paper aims to solve the latency problem in unauthenticated Byzantine Fault - Tolerance (BFT) consensus protocols in partially synchronous environments. Specifically, the paper proposes a new protocol named TetraBFT, which achieves an efficient consensus mechanism without relying on public - key cryptography and ensures robustness against attackers with unlimited computing power.

#### Main problems and goals:

1. **Reduce latency**:
- Existing unauthenticated BFT consensus protocols (such as IT - HS) require 6 message latencies to reach consensus under ideal conditions. TetraBFT reduces this latency to 5 message latencies by optimizing the protocol design, thus improving efficiency.

2. **Resource optimization**:
- TetraBFT only requires a constant amount of local storage space, and the communication complexity is \(O(n^2)\), that is, the number of messages sent and received by each node is linear. This makes TetraBFT more feasible in large - scale systems.

3. **Optimistic responsiveness**:
- When the network is synchronous, TetraBFT can respond quickly and reach consensus without waiting for a fixed timeout. This feature is very important for practical applications because it can avoid performance bottlenecks caused by network asynchrony.

4. **Pipeline processing of multi - round consensus**:
- The paper further extends TetraBFT to enable it to support multi - round consensus (also known as state - machine replication). By introducing pipeline technology, TetraBFT can commit a new block within each message latency, theoretically achieving five times the throughput of single - round consensus.

#### Specific implementation methods:

- **Security analysis and verification**:
- The paper proves the security and correctness of TetraBFT through strict theoretical analysis and formal verification (using the TLA+ language and the Apalache model checker).

- **Protocol design**:
- The design of TetraBFT includes four key stages: proposal, voting, decision - making, and view change. Through these stages, the protocol ensures that all honest nodes will only vote for safe values, thus guaranteeing the security and liveness of the protocol.

In conclusion, the goal of this paper is to solve the problems of high latency, large resource consumption, and difficulty in applying to actual blockchain systems in existing unauthenticated BFT consensus protocols by proposing the TetraBFT protocol.