FCDSB: A Fog Computing Network Architecture Based on Dynamic Sharding Blockchain for Consumer Electronics in AIoT

The Artificial Intelligence of Things (AIoT) is revolutionizing the consumer electronics sector by enabling intelligent interaction and real-time adjustments in consumer electronics products to provide personalized services. Fog computing extends computational services to the network edge, further enhancing the potential of AIoT. However, since fog layer devices dedicate most resources to executing core programs and data transmission, supporting additional security and privacy protections becomes highly challenging. Sharding blockchain, with its unique data-sharing mechanism, facilitates information flow in an open and trustless network, allowing secure and efficient management of large-scale AIoT data. Nevertheless, existing sharding blockchain systems face scalability issues due to low throughput and are susceptible to security risks from malicious attacks, limiting their application in AIoT. To address these challenges, this paper proposes a fog computing network architecture based on dynamic sharding blockchain, called FCDSB. The scheme optimizes sharding strategies by establishing a node reputation mechanism to reduce the impact of malicious nodes. The level of malice in the network is assessed by analyzing consensus inconsistencies, which helps determine the sharding boundaries under the current network conditions. Furthermore, deep reinforcement learning (DRL) is utilized to adaptively optimize system throughput and security levels to meet the evolving network conditions and demands. Simulation results indicate that this scheme significantly improves the scalability of sharding blockchain in AIoT applications while maintaining a high level of security.