Integrating machine data into software applications is not a new challenge: HMI and Scada systems, for example, require access to this data, and traditional data integration solutions - appliances, gateways or software applications - have been meeting these requirements for many years. However, the IIoT and traditional applications differ considerably in some respects: in traditional applications, the IT side remains stable over the lifetime of the system. Low maintenance is crucial, and it is ideal for the user if the data integration components do not need to be touched for years after initial installation and commissioning. In addition, the IT application runs locally at the site, and it should therefore also be possible to operate the data integration autonomously at the site without any further dependencies.

Figure 1: Edge computing enables the decentralized processing of data, managed via a central platform.

© Softing

The Industrial Internet of Things is different: Here, IT is the driver of innovation and innovation cycles are short. There is a whole bouquet of applications that is constantly evolving. Based on an expandable, flexible architecture, users are looking for a manageable entry point into an IIoT solution that can be modified and expanded over time. Applications should be provided via a central platform and rolled out as easily as possible across multiple production sites.

In terms of collecting machine data, this means Users must deal with the fact that data integration evolves with the solution over the lifetime of the IIoT solution and is subject to regular configuration changes. This dynamic is linked to the use of a central platform.

This means that the question of operating concepts for data integration is becoming increasingly important: data integration can no longer be solved autonomously and in isolation at a single location. Instead, data integration must be integrated into processes for operating an IIoT solution in the interaction between the central platform and the location(Fig. 1).

The Industrial Edge

Alongside the question of operating concepts, users of IIoT solutions are confronted with the challenge of being able to process data on site in the plant, not just within the central platform. Both the OT and IT industries provide functions and services for this task that can be summarized under the keyword 'edge computing'. Edge computing enables the decentralized processing of data, managed via a central platform. In relation to the IIoT use case, it is appropriate to speak of an 'industrial edge'.

In terms of technology, virtualization and the support of Docker containers play a major role here. Most major cloud platforms - including Amazon AWS and Microsoft Azure - support Docker containers as components for edge computing. Tool environments are also available, such as Kubernetes, which allow centralized management of Docker containers independently of cloud platforms.

Figure 2: Functional components of an edge node

© Softing

On closer examination of this situation, it becomes clear that the functionality of data integration can be understood as part of an industrial edge. This understanding of the industrial edge and the connection with data integration was recently described in detail in a white paper from the Open Manufacturing Platform (OMP)(Figure 2). In such an architecture, all the tools and services of the central platform are available for managing decentralized data integration. Depending on the use case, it may even be the only functionality of the Industrial Edge. In any case, this opens up new possibilities for users to manage the data integration of their IIoT solution centrally and flexibly, adapted to their organization and processes. As part of an Industrial Edge, it thus becomes an essential building block for the efficient and scalable operation of an innovative architecture and IIoT solution.

The EdgeConnector

Softing responded to these technology and market trends and presented an initial prototype of the EdgeConnector Siemens at SPS 2018, in conjunction with Microsoft Azure. The EdgeConnector is a software module that can read process and machine data from Siemens controllers via proprietary interfaces and make it available to IT via standardized interfaces. As a Docker container, it can be easily managed via a central platform. The available commercial product not only supports OPC UA but also MQTT at the interface to IT.

With EdgeConnector 840D and EdgeConnector Modbus, two further Docker containers are now available. All EdgeConnector products have a built-in web interface for configuration as well as an API for configuration via third-party applications. Further container products are currently in development or planned on the roadmap, including an OPC UA aggregation server and modules for collecting device data for asset management and asset monitoring applications. Softing also has its own 'Multi Factory Device Management System' in its portfolio, which is tailored to store floor requirements. This allows customers to design and implement independent architectures and IIoT solutions between the store floor, edge and cloud.

Figure 3: Schematic architecture of an IIoT solution with EdgeConnector Siemens AWS IoT SiteWise.

© Softing

As an independent specialist in the field of data integration, Softing is also strategically well positioned to take on a role as a bridge builder between OT and IT. Many industrial customers lack experience in edge computing and do not have access to reference artifacts and best practice resources for edge computing and the Internet of Things. This is where Softing benefits from the close partnership with major cloud providers that it has maintained for years. With this in mind, Softing and Amazon Web Services have jointly developed an 'AWS Quick Start' for EdgeConnector Siemens and AWS IoT SiteWise, which AWS makes available via its website(Fig. 3).

Implementation with AWS Quick Start

The AWS Quick Start automates the deployment of EdgeConnector Siemens and AWS IoT SiteWise in the AWS Cloud. Data is generated with a simulated Siemens S7-1500 PLC, sent to AWS via EdgeConnector Siemens and visualized with AWS IoT SiteWise. The simulation of the data is a standard feature of the EdgeConnector Siemens product. The entire deployment on AWS takes no longer than ten minutes.

The EdgeConnector Siemens running in the AWS cloud behaves exactly as if it were used in production and connects a real Siemens PLC. Users who do not have access to a PLC can try out possible applications and gain experience with an industrial IoT solution scenario. Users with little experience of AWS technologies benefit from the automated deployment of AWS IoT SiteWise.

With the help of AWS CloudFormation templates, the setup can be done automatically. They make it possible to specify the definitions of the necessary cloud resources and services in order to fully build a software stack. The automation scripts are open source-based and can be modified for individual projects as required.

Figure 4: The AWS IoT SiteWise dashboard displays control data simulated by EdgeConnector Siemens.

© Softing

Users do not need a license for edgeConnector Siemens to operate it as part of the AWS Quick Start: EdgeConnector runs in 'demo mode', which means that the full functionality of the product is available for 72 hours. After that, the product must be restarted and will then run for a further 72 hours.

If a user wants to switch from the virtual deployment in the AWS cloud to a deployment of edgeConnector Siemens on an edge device and connect a real PLC, essential parts of the Quick Start scenario can be reused(Figure 4).

Possible next steps

The AWS Quick Start provides an automated deployment of an industrial IoT solution. Users and developers can quickly and easily gain experience with the integration of machine data in AWS applications in a simulated store floor environment. Large parts of the solution can be adopted unchanged if real controllers are to be connected in an automation network. Possible extensions to the currently available AWS Quick Start can be outlined as follows:

The author: Dr. Christopher Anhalt is Vice President Product Marketing at Softing Industrial Automation.

© Softing

Firstly, the virtual Quick Start scenario can be supplemented by the automated provision of additional software components, for example by integrating edgeConnector Modbus.

Secondly, the Quick Start can be extended by automated provision of the edge components on a real IPC and by connecting a real controller.

And thirdly, the integration with services of the AWS platform can be deepened, for example by passing on and using semantic information with the aim of automating the creation of asset models at cloud level. The first version of AWS Quickstart already provides users with an effective tool that helps them to close the gap between OT and IT.