Companies are increasingly leveraging the power of the Industrial Internet of Things (IIoT) to transfer process and device data to the cloud and use it to improve their operations and efficiency and reduce costs. The industrial edge is often the backbone for delivering IIoT solutions, offering connectivity, integration of IT with OT, and data management while extending the capabilities of the cloud on-premises. These functionalities complement the cloud capabilities of Amazon Web Services (AWS) and Microsoft Azure, which are front runners in providing components and services to develop end-to-end IIoT solutions, including the edge layer. Furthermore, open-source container tools such as Kubernetes are being discussed more and more in the context of the edge. In this blog, we explain the industrial edge and the industrial IoT architecture and we introduce you to three projects that leverage AWS, Microsoft Azure platforms and Kubernetes and are already deployed or still under discussion in an earlier project phase.
What Does Industrial Edge Mean?
Connecting Production Assets And the Cloud In An Industrial IoT Architecture
How To Leverage IIoT Architectures With AWS And Microsoft Azure
With Flexibility And Openness To Complete IoT Integration
The industrial edge facilitates the connectivity between machines and devices in production plants and the cloud or centralized platforms. This allows shopfloor data to be used by other applications to facilitate management, improve productivity and efficiencies, or reduce costs. In a nutshell, the industrial edge extends the functionalities of a centralized platform (on the cloud or not) to the production or process sites.
The basic characteristics of an industrial edge system are:
If you want to take a deeper dive into the functionality of the industrial edge, download our detailed article.
So how is the industrial edge used in an industrial IoT architecture or IoT stack, and how does it help to connect the shop floor level to the cloud?
An industrial IoT stack or industrial IoT architecture can be described in its basic form as a three-layer model, connecting the shop-floor-level to the cloud. The bottom "Production Asset Level" comprises the area of operational technology (OT), which includes production machines, controls and sensors, among other things. The second level represents the edge level. It bridges the gap between factory or processes and the centralized (remote or non-remote) infrastructure and services in the upper level. The edge collects and consolidates data from multiple sources at the bottom level. Finally, on the top level, a central system obtains information from the plant assets and processes, can relay feedback downstream and manages and ‘orchestrates’ the edge.
Based on this theoretical definition, a plethora of implementations are possible and flourishing today, including a variety of components and actors within these layers. There is no "one-size-fits-all" solution for implementing such an industrial IoT stack. It is about ecosystems, partnerships, and integrations. What makes it exciting is that all of this is being shaped now, that there is no status-quo which presents both challenges and opportunities.
A key commonality across all implementations is the required connectivity and translation of information between the production and process assets (OT) and the edge and cloud layers (IT). These are two separate worlds that do not speak the same language. Such conversion might already happen at the production or process level but is often only provided at the edge. Moreover, the edge also aggregates, filters, and manages information as required and incorporates other indispensable functionalities such as security. A key technology to deliver, orchestrate and operate such functionalities are Docker containers. These self-contained and modularized software components can be deployed managed and configured with ease remotely and are a fundamental building block for IIoT solution architectures.
Forefront players in the IoT space are AWS and Microsoft Azure. They offer the most comprehensive range for cloud and IoT infrastructure as a service (IaaS), platform as a service (PaaS), and software as a service (SaaS). These offerings can be used to either build a custom solution from scratch or to leverage pre-built applications that accelerate development, deployment, and time-to-value. Both companies are cloud-native providers. Their services are open and enable integrations with third-party applications. They provide the ability to tailor and customize individual IoT solutions. Companies across the globe are leveraging AWS, Microsoft Azure and the Industrial IoT stack architecture we just described as a blueprint to harness the power of industrial IoT and to make use of production data to improve productivity and decision-making or reduce costs.
Here are three examples of recent projects built on AWS and Microsoft Azure.
Our first architecture is based on AWS IoT Greengrass, which is the edge environment from AWS for simplified deployment, management, and operation of container-modules, and provisioning of essential functional blocks. Among others, AWS IoT Greengrass provides the tools to build an integration with the AWS cloud. Third-party containers such as edgeConnector or the new edgeAggregator from Softing (to be released in the first quarter of 2022) can be used to provision and manage the southbound connectivity with the shop floor and the data collected. In our example, a company in the process industry developed a centralized IoT platform with AWS that connects and monitors different locations worldwide, with plans to roll out to 200+ different sites. Conversion of machine data to OPC UA is handled at the shopfloor level with a fleet of locally managed gateways and is thus not part of the edge. The key functional requirement for the edge level here is to consolidate data from the tens of OPC UA servers on the gateways and harmonize, filter, and organize the data before it enters the central cloud platform via a customizable endpoint. A solution for the company would be to use edgeAggregator container for filtering and data aggregation which provides the necessary integration, configuration versatility, and security functionalities to fulfil the current needs and adapt to future requirements.
An increasingly discussed variant of the industrial IoT stack architecture provides for two levels at the edge. At the top, AWS IoT Greengrass runs the MQTT broker that sends data to AWS. At the bottom, Kubernetes, an extensively used open-source container orchestration system, operates the different container instances that deliver the southbound connectivity with the production assets. Kubernetes (and other similar tools) allow for a high level of customization and for self-developed container management systems. The drawback is the extra effort necessary to deploy and operate the containers, which solutions such as AWS IoT Greengrass don’t have. Kubernetes can also be used in a single edge layer architecture that includes containers for cloud connectivity. This approach minimizes cloud vendor lock-in at the edge, helping companies remain more flexible.
In our case, an automotive OEM uses Softing’s edgeConnector family orchestrated within Kubernetes, to collect data from PLCs and CNC machines and forward it via the MQTT protocol. This two-layer edge scheme provides the customer with extra versatility regarding the OT/IT integration (bottom layer), while simplifying the effort and overhead for reliable and secure data transfer to the cloud (upper layer).
Our third use case features an architecture implemented by a vendor of compression machines using Azure IoT edge as environment for the edge level. Data from Siemens PLCs is collected by Softing’s edgeConnector Siemens, mapped into an OPC UA structure and fed into the Azure OPC Publisher. The data can be seamlessly and securely transferred from edge to cloud, with Azure IoT Hub as receiving end point, to be stored or consumed by applications. Operating or troubleshooting the automation network of the compression machines on location is not an option in most deployments, and so the question of how to manage the edge layer, including connectivity, is critical for the compression machine vendor. With edgeConnector, the company collects and converts data on site. But it can also configure remotely how this is done and adapt other functionalities such as security features. Deployment and orchestration of edgeConnector Siemens is facilitated by Azure IoT Edge, simplifying the effort of managing containers. This implementation capitalizes on Azure and the Softing solution to create the complete industrial IoT stack with limited effort utilizing off-the-shelf tools.
In conclusion, AWS and Azure provide flexible tools to produce complete IoT integrations adaptable to the diverse sets of requirements and use cases spurred by the digital transformation. The openness of these platforms and integration with third parties is key to address the challenges and deliver well-rounded systems. Kubernetes and other container orchestration systems will also play a central role in projects where a high level of customization of the edge operation functions is required. As a strategic partner of AWS and Azure for connectivity and data management at the edge, Softing uniquely leverages the benefits from these cloud platforms and seamlessly integrates with them in different scenarios – either directly deployed with their edge infrastructure services, or in open-source container management systems. Softing products are the cornerstone to provision machine connectivity, data translation and data management functions when working with AWS, Azure and Kubernetes to deliver reliable IIoT solutions.