Research On Time Sensitive Edge Computing Network Technology

With the extension of informatization in industrial production and the vigorous development of big data and artificial intelligence,the Industrial Internet has attracted more and more attention as a new type of infrastructure.However,in industrial production,data islands still exist,and manufacturers still need to improve the level of equipment informatization and networking.Time-Sensitive Networking(TSN)is a new-generation Ethernet standard that supports deterministic communication.It has attracted widespread attention in the industry due to its excellent real-time and deterministic capabilities,and has become the best choice for future industrial networks.Edge computing is a new computing paradigm close to the source of things or data,which reduces the risk of data leakage while responding quickly to data analysis requests.The combination of TSN and edge computing is not only conducive to solving the problem of data islands commonly existing in industrial applications,promoting IT/OT integration,and improving the certainty of network transmission and edge computing.This paper studies timesensitive edge computing network technology,mainly including the following work:(1)Investigate the related research in the field of TSN and edge computing.According to the structural and technical characteristics of TSN and edge computing,the edge server supports the computing tasks in TSN as the starting point to realize the integration of TSN and edge computing.Based on this,a time-sensitive edge computing network scheme is designed to provide deterministic transmission capability through precise gating of TSN network nodes in a time-synchronized environment.(2)Based on the above scheme,a deterministic edge computing method is proposed.The method utilizes the time-sensitive edge computing network to realize the deterministic transmission of computing task information,combines the interrupt threading of the real-time operating system,high-precision clock and other technologies to realize the deterministic computing of the task,and designs a scalable computing task load balancing module to calculate the collaborative task and information transmission.Then in the computationally intensive scenario of industrial production,the effectiveness of the load balancing module is simulated and verified by applying four load balancing algorithms.The simulation results show that after applying the load balancing algorithm,the maximum computing delay can be reduced by more than 50%,which verifies the effectiveness of the load balancing module.(3)In order to verify the deterministic transmission and deterministic computing capabilities of the above schemes and methods,this paper builds a time-sensitive edge computing network test bed,and combines the two scenarios of industrial control and industrial data analysis to conduct deterministic transmission and deterministic computing.The experimental results show that in industrial control scenarios,the proposed scheme can reduce the transmission delay of industrial control instructions from 100μs to 10 μs,and the jitter range from about 10 μs to about 2 μs;in industrial data analysis scenarios,the method in this paper can reduce the longest running time of data analysis from more than 250 ms to less than 70 μs,and ensure the smooth operation of the program,verify the certainty of calculation,and prove that the scheme and method are effective.