Recently, rapid advances in information and communications technologies (ICTs) have improved the performance of software-defined industrial Internet of things (SDIIoT). Due to a variety of data, flows and smart devices in SDIIoT, a distributed SDN control plane is necessary. However, how to achieve consensus among controllers efficiently and safely is an intractable problem. In this paper, we use a permissioned blockchain approach to reach consensus in distributed SDIIoT. Since lots of data needs to be synchronized, and the throughput of traditional Byzantine fault tolerance protocol used in permissioned blockchain is limited by many aspects, i.e., the trust features of nodes in blockchain, the trust features of controllers in distributed SDIIoT, and the computing capacity of blockchain, we jointly consider these aspects to improve the throughput of permissioned blockchain. Accordingly, we formulate view change, access selection, and computing resources allocation as a joint optimization problem. This joint problem is highly dimensional, and is hard to be solved by traditional methods. Therefore, we propose a dueling deep Q-learning approach to address the problem. Simulation results show the effectiveness of our proposed scheme.

Additional Metadata
Keywords Byzantine fault tolerance, dueling deep Q-learning, multiple controllers, permissioned blockchain, Software defined Industrial Internet of things
Persistent URL dx.doi.org/10.1109/GLOCOMW.2018.8644520
Conference 2018 IEEE Globecom Workshops, GC Wkshps 2018
Citation
Qiu, C. (Chao), Yu, F.R, Xu, F. (Fangmin), Yao, H. (Haipeng), & Zhao, C. (Chenglin). (2019). Permissioned Blockchain-Based Distributed Software-Defined Industrial Internet of Things. In 2018 IEEE Globecom Workshops, GC Wkshps 2018 - Proceedings. doi:10.1109/GLOCOMW.2018.8644520