Research On Fog Network Optimization Strategy For Industrial Internet Of Things

With the rapid development of the Industrial Internet of Things(IIoT),various sensors and terminal devices generate a large amount of real-time data,and traditional cloud computing has been difficult to meet the real-time requirements of data processing in industrial environments.Fog computing is located between cloud computing and terminal equipment,providing distributed computing,networking,and storage resources closer to edge devices than cloud computing,and therefore has better communication latency performance.For real-time processing of big data in industrial environments,the bandwidth of cloud computing is insufficient,and the computing power of a single fog node is limited.Based on Cloud/Fog Hybrid Computing Network(CFHCN),this paper studies the data distribution strategy of nodes and the online construction of fog network.To solve the problem of different computing power and communication capabilities of heterogeneous nodes in CFHCN,this paper proposes a Dynamic Equivalent Delay Optimization(DEDO)algorithm.The algorithm eliminates the invalid nodes in the fog network iteratively,and transforms the problem into a convex optimization problem,and finally obtains the data distribution ratio that minimizes the data processing delay.At the same time,the DEDO algorithm uses dynamic network measurement technology to dynamically adjust the data distribution ratio,which reduces the impact on the performance of the fog network when the network channel is disturbed in the industrial environment.Simulation results show that the DEDO algorithm has lower data processing delay than the Round Robin algorithm and the CPSO-LB algorithm.To solve the problem that node parameter information cannot be predicted due to the dynamic change of nodes in CFHCN,this paper proposes an Adaptive Fog Network Update(AFNU)algorithm to model adaptive fog network update into dynamic programming problem.According to the current fog network performance and the parameters of the new node,the algorithm selects a node that can reduce the minimum delay of data processing in the fog network system to join the fog network.The simulation results show that the fog network constructed by AFNU algorithm has the same system minimum delay as the fog network constructed under the condition of predicting all node parameters in CFHCN.