FDC: A Secure Federated Deep Learning Mechanism for Data Collaborations in the Internet of Things

TLDR

The explosive growth of IoT network data has increased demand for multiparty computation, and data are becoming valuable virtual assets for sharing and usage. This paper proposes FDC, a secure data collaboration framework that enables multiparty data computation in IoT without transmitting data outside private data centers. FDC combines a public data center, a private data center for governance, and blockchain to secure data usage and transmissions, and is validated in a real IoT scenario.

Abstract

With the explosive network data due to the advanced development of the Internet of Things (IoT), the demand for multiparty computation is increasing. In addition, with the advent of future digital society, data have been gradually evolving into an effective virtual asset for sharing and usage. With the nature of the sensitivity, massiveness, fragmentation, and security of multiparty data computation in the IoT environment, we propose a secure data collaboration framework (FDC) based on federated deep-learning technology. The proposed framework can realize the secure collaboration of multiparty data computation on the premise that the data do not need to be transmitted out of their private data center. This framework is empowered by public data center, private data center, and the blockchain technology. The private data center is responsible for data governance, data registration, and data management. The public data center is used for multiparty secure computation. The blockchain paradigm is responsible for ensuring secure data usage and transmissions. A real IoT scenario is used to validate the effectiveness of the proposed framework.