China Mobile and Huawei Carry Out a PoC Project in Edge Computing-Enabled Smart Factories

At Mobile World Congress Shanghai in July 2019, I had the privilege of chairing the Multi-access Edge Computing (MEC) session. It proved to be a particularly popular session, especially the GSMA’s keynote presentation on its “Edge Computing Smart Factory Proof of Concept (PoC)” project initiative. The GSMA’s Internet of Things (IoT) Program Steering Group (PSG) had invited more than 10 Communication Service Providers (CSPs) to put forward proposals for the PoC Project. Ultimately, the GSMA selected China Mobile’s and Huawei’s “Smart Factory Project” for the Edge Computing PoC project.

China Mobile and Huawei since invited industrial manufacturers to participate as project implementation partners. Haier (headquartered in Qingdao, China) hosted the smart factory PoC project and Hangzhou Mstar agreed to participate as Haier’s machine vision solutions application partner.

In mid-November 2019, the PoC project managers provided an update: a 5G MEC-enabled network was installed in Haier’s refrigerator factory that supported a machine vision application pilot. The PoC project managers intend to test out additional application scenarios in its next phase. China Mobile and Huawei consider machine vision, manufacturing pipeline, and video surveillance to be ideal applications for MEC campus scenarios.

At MWC Shanghai, China Mobile also outlined its “5G+AICDE” vision. In addition to being based on 5G, the vision integrates Artificial Intelligence (AI), IoT, cloud computing, big data, and edge computing into China Mobile’s approach.

Machine Vision was selected for the first PoC test case, as it leveraged several aspects of China Mobile’s “5G+AICDE” ambitions. While machine vision is a mature technology used in manufacturing, each machine vision device needs to be supported by a dedicated computer, which takes up space and Information Technology (IT) resources. Furthermore, each machine vision device is essentially isolated, and therefore it can be labor-intensive to configure each machine vision device and it can lead to data silos, complex IT maintenance, time consuming software upgrades, and reduced deployment rates.

A MEC-supported machine vision solution can be deployed centrally and serve all the machine vision devices with high bandwidth, high-performance computing, and low latency. Since the solution is cloud-based, it is relatively straightforward to implement alternative scenarios, such as Optical Character Recognition (OCR), or assess the current product assembly status quo. Furthermore, the cloud-based MEC solution can support deep learning and self-optimization.

Clearly, the GSMA Edge Computing Smart Factory PoC project is still a work in progress. There is potential in MEC supporting the needs of smart manufacturing along with other enterprise scenarios, but the PoC project managers stressed that MEC must deliver “simplicity, collaboration and efficiency.” This potential is underscored by ABI Research’s own research. In 2018, ABI Research estimated smart manufacturing solutions generated US$4.5 billion in revenue. By 2026, it is expected to grow to US$51.5 billion. A simplified architecture of centralized Operations and Maintenance (O&M) and maintenance-free edge nodes provides plug-and-play deployment and simplified O&M. High performance computing enables integration between multiple apps in each of the edge nodes.

This integration of applications means there is a strong need for collaboration between a number of upstream and downstream ecosystem players. Many of these MEC ecosystems are still very immature, with technology platforms that need further development. However, the main barrier has been developing the business models to support the MEC vision. This is the second goal of the GSMA Edge Computing Smart Factory PoC project phase two. The PoC project stakeholders recognize the need to further develop the relevant business models to support the various MEC-enabled use cases.

The GSMA and its stakeholders are not the only ones evaluating MEC business models, the underlying technology, and the software. In addition to the GSMA, the European Telecommunications Standards Institute (ETSI), the Edge Computing Association and Open Edge Computing Initiative, have also been evaluating MEC business models and testing PoCs.

Furthermore, a number of CSPs are keen to innovate in the MEC sector. In India, Vodafone Idea is leveraging Red Hat technologies to launch intelligent, DevOps-ready distributed edge computing platform in over 100 locations. Other CSPs have taken a consumer or prosumer angle to their MEC trails and deployments. Verizon Wireless is collaborating with video game publisher Bethesda Softworks and Amazon Web Services (AWS) Wavelength to deliver frictionless, ultra-low latency experiences for millions of gamers. Verizon Wireless is also leveraging AWS Wavelength to deliver in-stadium video enhancements for fans.

These PoC trials are invaluable, and they should serve to jumpstart the MEC innovation wave, but infrastructure vendors, CSPs, and enterprise customers will need to prototype, evaluate the results, and then iterate. Some enterprise customers will see the first mover advantage and align solution providers that have the most cost competitive solution and the most versatile MEC app ecosystem and price. However, ABI Research does believe there is an Industry 4.0 break-out opportunity for such first mover opportunists.