Managing the industrial edge: challenges, approaches and solutions

The use of edge computing in industrial environments is not new. The use of Programmable Logic Controllers (PLCs), microcontrollers, servers, and industrial PCs for local data processing or even micro data centers are edge technologies and have been used for decades in factory and automation systems.

Take our survey: Speak up, add your voice to our 2024 State of Initiative Report

What has changed is the emergence of the IoT, Industry 4.0, advances in AI and machine learning and new standards focused on plant and factory automation. These innovations are converging to support the digital transformation of operations across industrial manufacturing. In this context, a new generation of edge computing solutions is already revolutionizing the industrial plant and automation markets.

The business value that new edge computing solutions can bring to industrial markets is becoming widely recognized and organizations are starting to deploy at scale. However, challenges remain due to the heterogeneous nature of the industrial edge. These include:

Supporting the wide range of devices that these applications must run on, such as PLCs, MCUs, IPCs, IoT Gateways, and servers.
Configuring the applications to communicate with the many different operational technology (OT) endpoints.
Monitoring and updating a deployed system during its lifetime.
Ensuring the security of the end-to-end system, which may rely on the integration between OT and IT/cloud environments.

The role of edge management platforms

Trying to manage this complexity at scale in a system that could consist of hundreds or thousands of nodes (e.g. an edge device such as an IoT Gateway or IPC) and their applications with a “man-in-the-loop” is not practical.

To address these challenges, a new generation of edge management platform is emerging, focused on the specific needs of modern industrial edge systems and is becoming a key factor in enabling efficient, secure and scalable operations.

Podcast from IFS Unleashed: Focus on small, medium-size manufacturers

One of the key aspects of these new dedicated edge management solutions is the ability to centralize the management of an individual system or a deployment composed of multiple systems. This has significant benefits including:

Visibility and control: Provides a single pane of glass to monitor and control all edge devices and applications across different locations. This visibility is essential for ensuring that all devices function correctly and efficiently.
Efficiency and productivity: Allowing for remote management and maintenance of edge devices and their applications and reducing the need for on-site visits. This can significantly reduce operational costs and time.
Scalability: Ability to easily scale to accommodate new devices and applications. This flexibility is vital as industrial environments grow and evolve.
Security: Enforces consistent security policies across all edge devices and applications, helping to protect against cyber threats. This also helps to ensure that all devices are updated with the latest security patches and configurations.
Data integration: Facilitates the integration of data from various sources, enabling more comprehensive analytics and better decision-making.

Exploring edge management approaches

Different approaches can be taken to centralizing the management of an industrial system, these can be categorized as follows:

Cloud-based solutions: These solutions offer scalability and ease of management through cloud platforms.
On-premises solutions: These are suitable for organizations that require control over their data and infrastructure.
Hybrid solutions: These combine the benefits of both cloud and on-premises deployments, offering flexibility and control.

Unsurprisingly, solution providers in this space, especially major cloud vendors, are addressing this market with products that leverage cloud-native application management/orchestration technologies such as Kubernetes.

This approach has been only partially successful as Kubernetes supports the orchestration of containerized workloads and does not support the deployment and management of native binary applications, which are still used extensively in industrial systems. Additionally, due to its memory requirements, Kubernetes is not suitable for use with the smaller form factor edge devices (e.g., PLC or MCU) typically found in industrial systems.

The challenge is much broader in scope than just edge application orchestration though. An end-to-end edge management solution must address the management of the nodes on which the applications will run and the OT devices/sensors with which they will communicate.

Managing nodes and applications in industrial edge systems

At the node level, the scope of edge management tasks must include the ability to:

Register edge nodes securely and manage the installation and updates to any required software such as operating system, firmware or supporting drivers.
Monitor and manage the nodes throughout the lifetime of the system.
Easily debug any node-level problem.

Correspondingly at the application level, users require the ability to:

Remotely orchestrate the deployment of applications across a fleet of nodes.
Support different types of application workloads such as containers, binaries and scripts.
Automate the deployment of application updates throughout the lifetime of a system.
Monitor applications and address deployment issues.
Easily onboard and configure devices/sensors or cloud endpoints used by applications.

Overcoming device and sensor integration challenges

Sensors and devices connected to edge nodes communicate in myriad data formats and protocols, which introduces additional complexity to deploy, orchestrate, onboard/provision and monitor such applications.

eHandbook: Edge Computing

At the edge, configuring these applications can be a significant challenge. Even if the same application is going to be deployed across many nodes, variations in the specifics of the connected devices and other endpoints may require individual configurations to be applied on a per-node basis. So, it’s important to select a solution that provides a scalable way to deploy and update application-specific configurations.

Choosing the right edge management solution

Many of the edge management platforms that you can choose from focus on specific aspects of managing your industrial edge systems, typically either application orchestration or device management.

This means that users may either need to integrate multiple technologies to support their end-to-end use cases or perhaps implement extensions of top-of-the-base platform capabilities, both of which come with additional costs.

In addition to the functional and technical considerations of selecting an edge management solution, organizations need to consider the commercial aspect of adopting this technology. This includes initial and ongoing licensing and maintenance costs and whether to choose a product based on an open source or proprietary code base.

Typically, the trade-off here is between the openness, protection against vendor lock-in, and lower costs associated with open source versus more comprehensive support and advanced features of a proprietary product that have a higher price tag.

In summary, a new generation of edge solutions based on the latest advances in IoT, Industry 4.0, and AI/ML technologies is generating significant value for industrial systems, a trend that will only accelerate in the future.

The deployment and management of these new technologies impose additional complexities on how systems are provisioned and managed, particularly at scale. Looking ahead, industry leaders will increasingly turn to dedicated edge management platform solutions to address this problem.