Edge computing involves moving data processing from centralized data centers to remote, distributed locations. This means that data can be consolidated and analyzed on site before being transferred to a data center. This eliminates the challenges of network connections in terms of bandwidth or latency. By reducing transmission delays, service outages can also be avoided. Edge computing is being used more and more across all industries. Many projects can currently be found in the telecommunications sector in particular in connection with the introduction of 5G. Service providers are modernizing their networks by moving their workloads and services from the data centers to the edge. Edge computing will also play an important role in the manufacturing industry in the future, with the provision of computing resources away from central data centers directly on a device such as a robot on the factory floor. When data is processed as close as possible to the production line, this is referred to as the factory edge in the case of factory halls.

On the way to the factory edge

Validated pattern deployment in the data center.

© Red Hat

The main drivers for Factory Edge in the manufacturing industry are primarily Industry 4.0 projects. This involves topics such as IoT, artificial intelligence, robotics, augmented and virtual reality, blockchain, 3D printing, digital twins and 5G. In order to successfully implement Industry 4.0 concepts, large volumes of data need to be analyzed quickly. This task is largely carried out directly on the production line with the networking of IT with the plants or control systems, i.e. at the factory edge.

The reference architecture for edge implementations in production environments.

© Red Hat

The manufacturing industry has always used modern technologies to drive innovation and optimize production processes. With the Factory Edge concept, it can now go a decisive step further by bringing processing performance closer to the data. Among other things, this can proactively detect potential errors on the assembly line, improve product quality and even reduce potential downtime through predictive maintenance. And even if the communication link with a central location is impaired, individual production systems can continue to operate.

Wide range of application scenarios

Specifically, a company with a factory edge infrastructure can optimally cover a wide range of application scenarios. These include, for example

• Asset management,
• predictive maintenance,
• condition monitoring,
• visual inspection and quality control during production,
• end-to-end traceability and data collection along the entire supply chain,
• production planning,
• individual production (batch size 1),
• data sharing services.

The topic of factory edge is currently gaining relevance across the board. For example, the OMP (Open Manufacturing Platform) initiative is working on the development of innovative Industry 4.0 solutions and IIoT (Industrial Internet of Things) applications. Founded by the BMW Group and Microsoft, OMP, of which Red Hat is also a member, aims to promote the digitalization of production through open source and open standards, particularly in the context of the factory edge and the Industrial Internet of Things (IIoT). Red Hat is actively involved in various working groups, including the topics of "IoT Connectivity" and "Manufacturing Reference Architecture". The task is to develop concepts for connecting IoT devices and machines to the cloud in order to digitalize production lines and optimize cloud-based IoT applications.

The open hybrid cloud infrastructure as a basis

It is clear that implementing the factory edge with a reorganization of processes and data usage requires a complex system landscape: with sensors, wireless gateways, operational dashboards, agile backend systems and tablets for the maintenance teams to monitor the entire system. In addition, the increasing networking and digitalization of production and logistics processes, i.e. the linking of IT with the systems or control systems on the production line, also leads to changing requirements for IT itself. Conventional IT architectures do not offer an adequate basis due to their limitations in terms of agility and flexibility, speed or scalability.

The realization of a modern production environment therefore also requires an adequate IT infrastructure - an infrastructure based on containers, Kubernetes, agile development, AI and ML as well as automation. All of these technologies are components of an open hybrid cloud or multi-cloud IT infrastructure that enables the provision of applications in short development cycles - in a dynamically scalable environment. A hybrid cloud provides a consistent platform for bare-metal and virtual environments as well as for private and public clouds. An open hybrid cloud platform must also support comprehensive edge implementations. Edge computing thus becomes an integral part of an open hybrid cloud strategy.

With a consistent platform across all environments, a unified development and operational experience can be established from the core to the edge. A hybrid cloud platform also supports the portability of applications. This means that specialist applications, which are often developed centrally, can be rolled out in different cloud environments right through to the edge components. From a technological perspective, container technology and Kubernetes as a container orchestration framework have also increasingly proven themselves in the edge and IoT environment.

The author: Wolfram Richter is Manager Chief Architects at Red Hat.

© Red Hat

There is no question that the implementation of innovative Industry 4.0 and IoT scenarios is becoming increasingly important in manufacturing. This is the only way to realize efficient and intelligent business processes. The key tools here will be factory edge implementations and - more broadly - open hybrid cloud architectures that support the uniform creation and execution of solutions in cloud, on-premises and edge environments.