Lightweight Group Authentication for Decentralized Edge Collaboration

Due to tedious interactions and intensive computations, existing group authentication methods suffer from long latency and high overhead for simultaneously identifying multiple devices associated with one common task. Considering the growing importance of decentralized edge collaboration, this article proposes a lightweight group authentication scheme by utilizing the collaboration process information. Specifically, the edge devices in the same task group generate tokens for authentication based on knowledge from previous rounds of learning-based collaboration. We develop both the random token generation and privacy-preserving token generation strategies for achieving quick authentication and security enhancement, respectively. The generated tokens are easy for the legitimate group members to verify but extremely difficult for attackers to predict. A group authentication protocol is proposed for edge devices to verify the others' identities based on their tokens simultaneously. The proposed scheme achieves lightweight authentication without extra security information in addition to generating and distributing any key (secret), resulting in both enhanced security and decreased cost. More importantly, the proposed scheme enhances security and provides continuous protections for the decentralized edge collaboration by utilizing its natural update of authentication knowledge before the next round of collaboration begins. Our simulation study considers an example of decentralized edge collaboration and verifies the proposed scheme.