SeCrowd: Efficient secure interactive crowdsourcing via permission-based signatures

Abstract

Crowdsourcing is a powerful paradigm to provide services by soliciting help from a pool of contributors. However, the security threats in crowdsourcing, especially in interactive crowdsourcing where contributors interact with each other for collaboration, are not well studied. External adversaries who may inject malicious results are the major concerns of all involved parties in crowdsourcing. In this paper, an efficient secure interactive crowdsourcing framework, called SeCrowd, is proposed to protect the authenticity in the crowdsourcing platform against external adversaries. The framework is characterized by a recently proposed cryptographic primitive, called permission-based signature, which reveals the signer’s identity to permissive verifiers only, while safeguards it against others. In this way, contributors can identify each other based on their signed contributions, but their identities remain unknown to other users. Moreover, users can verify whether the signed contributions are made by legal contributors without being aware of the actual signer’s identities, which guarantees that the contributions cannot be tampered or forged. We prove that SeCrowd achieves unforgeability in the random oracle model. Comprehensive analysis and experiments showed that our framework is secure and highly efficient in terms of computation and communication cost.