Cloud SPN Computing and Network Convergence Verified by China Mobile, ZTE and Xingqi

• Deploying edge computing power at low costs based on the Cloud SPN computing and network convergence solution, and driving industrial manufacturers towards digital and intelligent production transformation
• Slicing private line + AI computing service blazing a trail for operators to transform from traditional pipe operation to new ToB service operation

ZTE Corporation (0763.HK / 000063.SZ), a global leading provider of information and communication technology solutions, together with China Mobile, and Changzhou Xingqi Technology, recently piloted AI industrial quality inspection based on the Cloud SPN computing and network convergence to deploy edge computing power at low costs and drive industrial manufacturers towards digital and intelligent production transformation.

In traditional industrial parks, there are many pain points in quality inspection. Some plants rely on manual inspection rather than intelligent inspection, resulting in a significant consumption of human resources. Some are deployed with expensive dedicated devices for automatic quality inspection, but the changing quality check demands cannot be met quickly due to a variety of samples and inflexible algorithms of dedicated devices. Others have low precision in local automatic check, necessitating real-time adjustments to system parameters, leading to high service costs.

To solve these problems, China Mobile and ZTE have jointly come up with the Cloud SPN computing and network convergence solution, which uses the native HEC (Hyper-Edge Computing) at the access layer in the existing SPN to offer the local computing resource pool. The edge-cloud synergy solution is used to integrate edge computing, cloud computing, and network capabilities to provide slicing private line + AI computing services for the ToB vertical industry, so that AI applications can be deployed quickly in industrial parks.

Guided by China Mobile and China Mobile Research Institute, China Mobile Jiangsu Branch and China Mobile Changzhou Branch jointly started the exploration of Cloud SPN computing and network convergence technology in early 2022, and started verification of computing power deployment and application in the existing network in 2023.

The access SPN device is equipped with the computing board to rapidly apply the HEC to the access-layer SPN, and the slicing private line is used to implement synergy of terminals, edges and clouds to deliver the slicing private line + native AI computing service to the industrial on-site network. On one hand, broad coverage and large networking of the SPN facilitate ubiquitous deployment of computing nodes, and computing resources can be scheduled flexibly between them. On the other hand, computing is integrated into the network to simplify the network hierarchy, reduce the overall power consumption, and save energy.

With the on-site HEC device (the ZXCTN 6180H integrated with computing boards) provided by ZTE, Xingqi deploys the AI machine vision quality inspection application to fulfill the task of the spray code check process of its eighth station. The HEC is interconnected with the PLC (Programmable Logic Controller) of the industrial camera production line to collect data via the camera, send the execution results to the PLC after local AI inference, and then control the robot arms for production operations. The verification result shows that the quality inspection precision can be increased to 99.99%, which is a great improvement in the efficiency and accuracy compared to the traditional manual inspection. Furthermore, the investment in quality inspection equipment is cut by 75%, and the spending in computing construction to support AI applications at local is cut down by more than 90%. In addition, an industrial quality inspection process based on AI model can be flexibly switched according to orders. Compared with the conventional manner, quality inspection sampling and inference efficiency is raised by five times, from 24 to 120 times per minute.

Starting from the machine vision AI industrial quality inspection scenario and going deeply into production application process, this pilot explored a general system solution for interconnection between the HEC device supporting the computing and network convergence and the traditional industrial PLC system, which can interconnect with the PLC from multiple vendors to solve problems of interconnection, adaptation and switching of traditional quality inspection equipment.

After two years’ pilot practice, China Mobile, ZTE and Xingqi have verified the technical feasibility of slicing private line + AI computing service, started deploying the hyper-edge computing and network convergence solution, and found the value-added business process and service process for commercial operation that can be promoted. These endeavors blaze a trail for operators to transform from traditional channel operation to new ToB service operation.