大阪大学　数理・データ科学教育研究センター
Center for Mathematical Modeling and Data Science,The University of Osaka

Sequential vs. Simultaneous Blockchain Consensus with Adversarial Attacks

Yunlong Liu (University of Queensland)

1. ホーム
2. セミナー・集中講義
3. セミナー
4. Sequential vs. Simultaneous Blockchain Consensus with Adversarial Attacks

UQ-Osaka Seminar on Financial Mathematics and Economics 第2回 (University of QueenslandとMMDS金融保険部門共催)

Sequential vs. Simultaneous Blockchain Consensus with Adversarial Attacks

Yunlong Liu (University of Queensland)

In the past few years, the successful launch of Tendermint in blockchains has brought attention to the decades-old theory of the Byzantine Generals’ Problem from the computer science literature. The solution to this problem is known as Byzantine Fault Tolerance (BFT) protocols. The idea behind BFT-based blockchains is to introduce a multi-round voting mechanism to enhance the protocol’s fault tolerance. While BFT-based blockchains analyze a voting mechanism in a sequential context, we study a voting game in simultaneous and sequential voting protocols, capturing the major features of BFT-based blockchains. We characterize all equilibria, and the findings are expected to provide guidance for the successful design of blockchains. We find that sequential voting consensus is better than simultaneous consensus if no adversary is present in the committee voting. However, if an adversary appears in the voting, the choice of voting protocol depends on the society’s preferences and the evaluation of the trade-offs.