An Improved Proof-of-Trust Consensus Algorithm for Credible Crowdsourcing Blockchain Services

In online crowdsourcing services, credible accountability mechanisms are crucial for guaranteeing a good interactive environment. However, the crowdsourcing systems are established in virtual environments, the identities of the participants are various and complicated, the systems could scarcely identify malicious nodes automatically. So it is very hard to preserve the complete evidence of malicious behaviors and investigate relevant legal responsibilities. Blockchain is regarded as a very promising solution to these problems because it possesses characteristics of decentration, non-modifiability and traceability. However, a main challenge is to design an applicable blockchain consensus algorithm which can reach an agreement on credibility of participants automatically, prevent transaction data from tampering, and trace to the source of malicious behaviors. In this paper, an improved Proof-of-Trust (PoT) consensus scheme is proposed with the underlying technology of blockchain, which is properly to the crowdsourcing service scenarios. Firstly, this PoT consensus selects nodes with high credibility using subjective logic reputation algorithm. Only selected nodes have the chance to generate blocks, participate in verification, and claim crowdsourcing tasks. Secondly, the choice scheme of generate-block nodes is further optimized through the unpredictability of timestamp and digital signature. Moreover, an incentive mechanism based on game theory is designed in this consensus. With this mechanism, candidate nodes prefer to give honest verification results rather than engage in collusion with malicious nodes. The analysis and simulation results demonstrate the effectiveness, feasibility and scalability of the proposed approach.