[KCC 2023] Survey on Asynchronous BFT consensus Algorithms for Scalable and Robust Blockchain
Author:
Donghyeon Ryu, Chanik Park in POSTECH
Journal/Conference:
KCC '23
Paper:
(eng)
(kor)
Abstract
Blockchain technology is being applied to various areas such as non-fungible tokens (NFTs), decentralized finance (DeFi), security tokens (STs), central bank digital currencies (CBDCs), and decentralized identity (DID) based on its decentralization, transparency, and high reliability. However, the consensus process in blockchain requires all nodes to participate, which poses scalability limitations. To address scalability issues, synchronous BFT consensus algorithms, primarily from the PBFT family, have been applied. However, synchronous BFT algorithms are vulnerable to Byzantine faults and suffer from a concentration of load on leader nodes. As a result, more robust asynchronous BFT algorithms have gained recent attention. This paper aims to introduce asynchronous BFT algorithms and provide insights for future research by examining representative studies.
Synchronous BFTs
Asynchronous BFTs
(1) ACS-based BFTs
(2) DAG-based BFTs
- HashGraph
- Aleph
- DAG-Rider
- Tusk
Hybrid BFTs
- Narwhal
- Dumbo-Bolt Transformer
- Ditto
- Bullshark
Conclusion
In this paper, we discussed the scalability issues and vulnerability to Byzantine faults in existing synchronous BFT consensus algorithms, and demonstrated how these can be addressed using asynchronous BFT consensus algorithms through recent research examples. We believe that this provides insights for practical BFT consensus algorithm research, considering the expected increase in demand for blockchain in the future.