Academic SeminarUnderstanding Blockchain Governance Decentralization: An Analysis and Agent-based Simulation of Algorithmic Governance in the Proof-of-Stake Consensus Mechanism
- 일시
- 2019-06-19 ~ 2019-06-19
IT경영 분야 세미나를 아래와 같이 개최하오니, 관심 있는 분들의
많은 참석 부탁 드립니다.
1. 일시: 2019년 6월 19일 (수), 16:00~17:20
2. 장소: 수펙스 경영관 401 강의실
3. 강사: 한정필 교수 (National University of Singapore)
4. 주제: Understanding Blockchain Governance Decentralization: An Analysis and Agent-based Simulation of Algorithmic Governance in the Proof-of-Stake Consensus Mechanism
5. 연구분야: IT경영 분야
* Lecture will be delivered in Korean.
* Seminar materials: Abstract
[Abstract]
Increasingly, algorithms do not only mediate or make decisions as such, but also assign decision rights. However, such algorithmic governance is little understood. The literature suggests algorithmic governance may be associated with two major limitations: unintended outcomes and opacity. We explore these two issues by studying governance in blockchain technology as a prominent example of algorithmic governance. Blockchains, best known as the underlying technology of cryptocurrencies and several other major Fintech innovations, are distributed databases of transactions functioning without a central operator. They are shared by a number of participants called nodes, with data stored in batches as blocks. To discourage the creation of alternative blockchains, consensus mechanisms incentivize agreement among the nodes by assigning the right to append a new block to one selected node, the validator node. Decentralization in consensus mechanisms is a critical aspect, giving rise to the main value proposition of blockchains, that is, tamper-resistant data storage. Recent attacks illustrate that decentralization cannot be taken for granted. We study how key blockchain parameters affect the degree of decentralization, thereby opening the black box of how algorithmic governance may lead to unintended consequences. We do so by analyzing and simulating an implementation of proof-of-stake, a salient consensus mechanism. We use agent-based modeling to identify behaviors of the consensus mechanism under a range of parameter manipulations. The results suggest that a high number of initial potential validator nodes, large transactions, and a high number of transactions increase the degree of decentralization, whereas high transaction fees and high growth in the number of potential validator nodes lower the degree of decentralization. Our findings can be applied in understanding the mechanisms that lead to decentralization in blockchain governance as well as in designing blockchains that are prone to decentralization and therefore more secure.