# The latest gossip on BFT consensus

**Authors:** Ethan Buchman, Jae Kwon, Zarko Milosevic  
**arXiv:** [1807.04938](https://arxiv.org/abs/1807.04938)  
**Date:** 2018-07-13  
**Summary page:** https://www.summarizepaper.com/en/paper/1807.04938v1/

## Key points

- Tendermint is a new protocol for ordering events in a distributed network under adversarial conditions.
- It builds upon classic academic work on Byzantine Fault Tolerant (BFT) consensus and introduces a simplified design for the BFT algorithm.
- The protocol utilizes a peer-to-peer gossip protocol among nodes for efficient communication and coordination.
- Advanced cryptography techniques, such as Boneh-Lynn-Shacham (BLS) signatures, are employed to enhance security and reliability.
- The system model considers processes that communicate through message exchanges, with some processes being correct and others faulty.
- A partially synchronous system model is defined, ensuring reliable message delivery after a global stabilization time (GST).
- Public-key cryptography prevents impersonation attacks throughout the execution of the system.
- Tendermint's state machine replication mechanism ensures consistency across all replicas in the network.
- Transactions are proposed, ordered within blocks, and committed to the replicated state machine.
- Tendermint offers an innovative approach to achieving BFT consensus in distributed networks while maintaining security and efficiency.

## Summary

The paper presents Tendermint, a new protocol for ordering events in a distributed network under adversarial conditions. This problem, also known as Byzantine Fault Tolerant (BFT) consensus or atomic broadcast, has gained significant attention in recent years due to the success of blockchain-based digital currencies like Bitcoin and Ethereum. These cryptocurrencies have successfully solved the problem of achieving consensus in a public setting without relying on a central authority. Tendermint builds upon classic academic work on BFT consensus and introduces a simplified design for the BFT algorithm. The protocol achieves this by utilizing a peer-to-peer gossip protocol among nodes. This approach allows for efficient communication and coordination between nodes in the network. The paper provides additional context by discussing the decentralization aspect of BFT consensus algorithms. It highlights the use of advanced cryptography techniques, such as Boneh-Lynn-Shacham (BLS) signatures, which are employed in protocols like SBFT to enhance security and reliability. In terms of the system model, the paper considers a network of processes that communicate through message exchanges. Processes can be either correct or faulty, with faulty processes exhibiting arbitrary behavior. The model assumes that each process has some voting power and is connected to a subset of other processes called peers. Communication between processes is facilitated using a gossip protocol. The paper also defines the partially synchronous system model, which introduces bounds on communication delays and ensures reliable message delivery among correct processes after a global stabilization time (GST). The use of public-key cryptography prevents impersonation attacks throughout the execution of the system. Furthermore, the paper outlines Tendermint's state machine replication mechanism, which ensures consistency across all replicas in the network. It describes how transactions are proposed and ordered within blocks before being committed to the replicated state machine. Overall, Tendermint offers an innovative approach to achieving BFT consensus in distributed networks under adversarial conditions. Its reliance on peer-to-peer gossip protocols simplifies the design while maintaining security and efficiency. The paper provides a comprehensive overview of the protocol, including its system model, cryptographic techniques, and state machine replication mechanism. This research contributes to the ongoing development of decentralized systems and has implications for various applications beyond blockchain-based digital currencies such as deployment and management in distributed networks.