Solution Probing Attack Against Blockchain-based Quality-aware Crowdsourcing Platforms

Conventional crowdsourcing platforms primarily rely on a central server as the broker for information exchange. Although many efforts have been made, centralized platforms are still vulnerable to underlying security issues such as the trusted central server and single-point failure. Fortunately, blockchain has emerged as an alternative infrastructure for building crowd-sourcing platforms. Many excellent designs of decentralized crowdsourcing platforms (DCSPs) built atop blockchain have been proposed recently. Benefiting from blockchain, DCSPs can provide fascinating features, like tampering resistance and anonymity. However, new security issues keep cropping up. While existing studies have proved that DCSPs are vulnerable to various attacks, our study has identified a new attack named solution probing attack against DCSPs. The adversary of solution probing attack can take advantage of the anonymity of DCSPs to probe valid solutions using a generative model. Due to the transparency of blockchain transactions, our probing attack is effective even if the solutions are encrypted. We evaluate the probing attack on large-scale crowdsourcing tasks. Experimental results show that the adversary is capable of rewarding fraud without any meaningful contribution to the crowdsourcing task. To protect DCSPs from the probing attack, we propose a defense solution based on a mix-and-match strategy. Defense evaluation results show that our solution can defeat the probing attack effectively.