Key points are not available for this paper at this time.
Fast-growing smart city applications, such as smart delivery, smart community, and smart health, are generating big data that are widely distributed on the internet. IoT (Internet of Things) systems are at the centre of smart city applications, as traditional cloud computing is insufficient for satisfying the critical requirements of smart IoT systems. Due to the nature of smart city applications, massive IoT data may contain sensitive information; hence, various privacy-preserving methods, such as anonymity, federated learning, and homomorphic encryption, have been utilised over the years. Furthermore, limited concern has been given to the resource consumption for data privacy-preserving in edge computing environments, which are resource-constrained when compared with cloud data centres. In particular, differential privacy (DP) has been an effective privacy-preserving method in the edge computing environment. However, there is no dedicated study on DP technology with a focus on smart city applications in the edge computing environment. To fill this gap, this paper provides a comprehensive study on DP in edge computing-based smart city applications, covering various aspects, such as privacy models, research methods, mechanisms, and applications. Our study focuses on five areas of data privacy, including data transmitting privacy, data processing privacy, data model training privacy, data publishing privacy, and location privacy. In addition, we investigate many potential applications of DP in smart city application scenarios. Finally, future directions of DP in edge computing are envisaged. We hope this study can be a useful roadmap for researchers and practitioners in edge computing enable smart city applications.
Yao et al. (Wed,) studied this question.