Performance Optimization Blockchain-Enabled Fog Computing with Deep Reinforcement Learning

Abstract

With the rapid development of Internet of Things technology and cloud computing technology, Fog Computing emerges with the demand of distributed and decentralized. Blockchain is considered to be an advanced technology, which enables Fog Computing data to be stored and shared safely, greatly improving the security and privacy of Fog Computing. However, the Blockchain technology currently applied in Fog Computing still has some significant disadvantages, one of which is the low throughput performance. In order to meet the demand of improving throughput, this paper proposes a Blockchain-Enabled Fog Computing performance optimization framework based on Deep Reinforcement Learning (DRL), which can dynamically adjust according to the system state to improve throughput. Dueling Deep Reinforcement Learning is adopted to obtain the optimal configuration, which can be dynamically adjusted according to Fog node status, network bandwidth and user's quality of service (QoS) requirements, including dynamic selection of block producers, block size and network bandwidth allocation. Simulation results show that the proposed scheme can effectively improve throughput performance.