Analysis of reputation-based mining paradigm under dishonest mining attacks

Abstract

Since the introduction of Bitcoin, numerous studies on Bitcoin mining attacks have been conducted, and as a result, many countermeasures to these attacks have been proposed. The reputation-based mining paradigm is a comprehensive countermeasure solution to this problem with the goal of regulating the mining process and preventing mining attacks. This is accomplished by incentivizing miners to avoid dishonest mining strategies using reward and punishment mechanisms. This model was validated solely based on game theoretical analyses, and the real-world implications of this model are not known due to the lack of empirical data. To shed light on this issue, we designed a simulated mining platform to examine the effectiveness of the reputation-based mining paradigm through data analysis. We implemented block withholding attacks in our simulation and ran the following three scenarios: Reputation mode, non-reputation mode, and no attack mode. By comparing the results from these three scenarios, interestingly, we found that the reputation-based mining paradigm decreases the number of block withholding attacks, and as a result, the actual revenue of individual miners becomes closer to their theoretical expected revenue. In addition, we observed that the confidence interval test can effectively detect block withholding attacks; however, the test also results in a small number of false positive cases. Since the effectiveness of the reputation-based model relies on attack detection, further research is needed to investigate the effect of this model on other dishonest mining strategies.