The AI-driven Permissioned Blockchain System for the Services of Future Internet of Things
Dajun Zhang, Wei Shi · 2023
Recently, the Internet of Things (IoT) and blockchain technology have attracted more and more attention. However, given the characteristics of the current blockchain itself, many important issues prevent it from becoming a general platform for IoT to deploy large-scale services. The main problem is how to ensure the scalability of the blockchain system in the dynamically changing IoT scene. Despite numerous studies conducted to address this issue, they failed to account for the dynamic changes in IoT nodes, specifically the addition and deletion of nodes. Instead, they simply opted for a consensus protocol as the presumed optimal solution. In this paper, we employ a permissioned blockchain framework designed for current IoT services. In response to the Quality of Service (QoS) requirements of different IoT nodes, we propose different consensus protocols for different service qualities in the existing IoT architecture, aiming to enhance the adaptability of the blockchain to the different needs of users in the IoT system. First, we quantified several permissioned consensus protocols. Furthermore, IoT services require the selection of block producers possessing significant computing resources, along with the dynamic allocation of network bandwidth specifically for the blockchain. We present an approach where we formulate the joint optimization problem of the selection of consensus mechanism and block producer, and the allocation of available bandwidth. We employ the Dueling Deep Q-network (DDQN) method as a solution for this problem.