YAC: BFT Consensus Algorithm for Blockchain

TL;DR

YAC is a practical Byzantine fault tolerant consensus algorithm designed for blockchain systems, improving message efficiency and leadership issues, with proven safety, liveness, and scalability to dozens of peers.

Contribution

The paper introduces YAC, a new consensus algorithm that addresses inefficiencies and leadership problems in classical BFT algorithms, with formal proofs and empirical scalability results.

Findings

YAC ensures safety and liveness in decentralized systems.

YAC scales to dozens of validating peers.

YAC is implemented in Hyperledger Iroha.

Abstract

Consensus in decentralized systems that asynchronously receive events and which are subject to Byzantine faults is a common problem with many real-life applications. Advances in decentralized systems, such as distributed ledger (i.e., blockchain) technology, has only increased the importance of finding performant and secure solutions to consensus of state machine replication in decentralized systems. YAC is a practical decentralized consensus algorithm, that solves the problems of inefficient message passing and strong leaders that occur in classical Byzantine fault tolerant consensus algorithms. The algorithm is open source and currently is used to provide Byzantine fault tolerant consensus for the Hyperledger Iroha blockchain project. We provide proofs of safety and liveness, as well as empirical results showing that our algorithm can scale to dozens of validating peers.