5G Mobile Edge Computing will empower OT in manufacturing

AT&T announced that it will embrace edge computing in its 5G network, in the form of Multi-Access Edge Computing (MEC – previously referred to as Mobile Edge Computing). MEC brings cloud computing from the network core, at the data centers, to the network ‘edge,’ at the central offices, cell towers and small cells. AT&T will equip its edge locations with graphic processing chips and computers to handle and process the massive amounts of data that both industrial and consumer Internet of Things (IoT) applications will generate. 5G MEC will provide latencies lower than 10ms and even as low as 1ms that allow near instant data processing and analytics, [1] compared to the average 79ms latency today’s LTE networks offer.

Low latency with 5G MEC means that data is sent from a connected sensor to the network edge, processed and sent back in milliseconds. That allows operations of all kinds to offload some of the computing power and still get instant results. Data-heavy applications that need near-instant data processing would no longer need to depend on only the data that is gathered at the source. Instead, a machine’s sensors can capture data and transmit it to the edge of the network, where it will be processed and enriched with updated information, feed machine learning models and sent back to the source for use and action, all in a few milliseconds.

For industrial applications, this has strong implications for predictive maintenance and collaborative robotics. Sensor data can be verified at the edge of the network, and digital twins can instantly update and efficiently trigger real-time insights with more security and privacy because the data is stored and processed more locally. Machine learning and OTA updates can download without delay. Collaborative robots on assembly lines can do the same, and by aggregating data at the edge will learn how to work with each other and with humans more quickly, safely and efficiently.

By easing processing and bandwidth requirements at the edge of the network, OT can more easily integrate with IT systems. Machines automatically update or engineers on the floor get instant actionable analytics. MEC also minimizes the back-and-forth on equipment testing because rather than running a test and waiting for the results, engineers can view instant measurements without having to rely on local computing.

For monitoring and diagnosis, shop-floor workers could receive near-instant alerts on whatever devices they carry when an issue arises. Or, if the machine learning and predictive maintenance works as intended, they will receive alerts before an issue arises. If the issue does not require immediate attention, or if down-time is inevitable, the system could schedule maintenance. Low latency makes all this possible because the digital twins will update in near real-time. This empowers the OT organization and fulfills the promise of the IIoT by sharing information without delay to minimize down-time.

MEC both makes connected machines more independent and more integrated into the IT system. They can process data near the source and get instant actionable analytics, but the data still goes to the cloud. CIOs will need to adapt their IT systems to the data and analytics taking place at the edge more so than operations constantly adapting to fit new IT systems. MEC distributes more of the computing to the edge, putting more importance on communications and a coordinated network.

See our report Edge Analytics in IoT: Supplier and Market Analysis for Competitive Differentiation for more information. 

[1] See http://www.prnewswire.com/news-releases/the-cloud-comes-to-you-att-to-power-self-driving-cars-arvr-and-other-future-5g-applications-through-edge-computing-300489292.html