Blockchain and Federated Deep Reinforcement Learning Based Secure Cloud-Edge-End Collaboration in Power IoT

• Published in: IEEE wireless communications Vol. 29; no. 2; pp. 84 - 91
• Main Authors: Zhang, Sunxuan, Wang, Zhao, Zhou, Zhenyu, Wang, Yang, Zhang, Hui, Zhang, Geng, Ding, Huixia, Mumtaz, Shahid, Guizani, Mohsen
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.04.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Blockchain; Blockchains; Cloud computing; Collaboration; Computation offloading; Computer architecture; Cryptography; Cybersecurity; Data retrieval; Delay; Delays; Internet of Things; Machine learning; Network latency; Optimization; Queueing; Reinforcement learning; Resource allocation; Simulation
• ISSN: 1536-1284; 1558-0687
• DOI: 10.1109/MWC.010.2100491

Cloud-edge-end collaboration enables harmonious and efficient resource allocation for the Power Internet of Things (PIoT). However, the security and complexity issues of computation offloading evolve into the main obstacles. In this article, we first propose a blockchain and AI-based secure cloud-edge-end collaboration PIoT (BASE-PIoT) architecture to ensure data security and intelligent computation offloading. Its advantages in flexible resource allocation, secure data sharing, and differentiated service guarantee are elaborated. Then the adaptability of three typical blockchains with PIoT is analyzed, and some typical application scenes of BASE-PIoT including computation offloading, energy scheduling, and access authentication are illustrated. Finally, we propose a blockchain-empowered federated deep actor-critic-based task offloading algorithm to address the secure and low-latency computation offloading problem. The coupling between the long-term security constraint and short-term queuing delay optimization is decoupled by using Lyapunov optimization. Numerical results verify its excellent performance in total queuing delay and consensus delay.