A Novel Privacy-Preserving Incentive Mechanism for Multi-Access Edge Computing

Abstract

Multi-access Edge Computing (MEC) has emerged as a promising solution for computation-intensive and latency-sensitive applications. Existing studies have often overlooked the critical aspect of users’ privacy, hindering users from offloading their computation. This paper proposes a novel privacy-preserving mechanism for a two-level auction game aimed at incentivizing cloudlets and users to engage in computation offloading while safeguarding users’ privacy. A many-to-many auction is designed between Data Center Operators (DCOs) and cloudlets to associate the cloudlets with the DCOs, where the perceivable privacy levels of users are parameterized as part of a DCO’s utility. A many-to-one user-DCO auction is also designed, leveraging differential privacy (DP) to protect the users’ private bid information. An exponential mechanism is developed, obfuscating intermediate reference prices disclosed during auctions by the DCOs, thereby safeguarding users’ valuations, bid prices, and bidding behaviors. We prove that the proposed approach can guarantee DP, truthfulness, and equilibriums. Simulations demonstrate the superiority of the privacy-preserving two-layer auction game in reducing time delay and energy consumption while protecting the privacy of the users, surpassing the benchmark. The proposed mechanism effectively incentivizes computation offloading, making it a compelling choice for facilitating computation-intensive tasks.