A Privacy-Preserving and Robust Reputation System Based on Blockchain
Shuai Sun, Yuan Liu, Guibing Guo · 2019
In the era of the Internet, agents behaving on behalf of human beings and smart devices are ubiquitously connected, and a well-designed reputation system is essential for them to exchange information efficiently. Reputation systems bear many challenging problems, such as data security, user privacy, as well as vulnerabilities against untruthful ratings (bad-mouthing and ballot-stuffing attacks), resulting in the ineffectiveness of the systems in differentiating honest users from malicious ones. Moreover, some users even hesitate to submit their truthful negative feedbacks due to the fear of the retaliation from recipient users. To address the above issues, we are to propose a privacy preserving and robust reputation system based on the blockchain. Specifically, first of all, we propose the reputation token concept and design a reputation token age based consensus mechanism to incentivize evaluators to actively participant the block mining process. Secondly, we consider the effects of the transaction price and evaluators' reputation in adjusting the weights of ratings so as to defect against ballot-stuffing and bad-mouthing attacks. Furthermore, we improve the malicious-k-shares protocol to enable the ratings to be aggregated in a privacy preserving way, where the rating privacy is not leaked in the whole process of rating storage and usage. Experiments based evaluations and discussions have demonstrated the effectiveness of the proposed reputation system.