The OPC UA standard is ideally suited for creating software definitions of industrial production facilities and their devices. Accordingly, the standard is used to ensure the interoperability of OT and IT: OPC UA servers make standards-compliant data available, while OPC UA clients access this data in standardized ways.
Information models are a core concept here, as the OPC UA information models describe the services and data offered by an OPC UA server. The OPC UA standard organizes information models into base models and optional model extensions. These extensions, which are specific to an industry or application domain, are termed “companion specifications.” The OPC UA standard also provides for manufacturer- and user-specific extensions to base models. These extensions may optionally use companion specifications as their starting point.
“OPC UA information models” is generally synonymous with all these variants and is therefore the term routinely used in this article.
Benefits of information models
As standardized interface definitions, information models reduce the effort required to integrate OPC UA server components and OPC UA client components. Depending on the end user, this offers various benefits, including the following:
- Companion specifications reduce system integration effort on the part of manufacturers and users, make it easier to replace devices from different vendors, and accelerate the development of IoT applications. Some of the larger and better-known companion specifications include PA-DIM for process automation and the Weihenstephan Standards for the food industry.
- In cases where no suitable companion specifications are available for a specific solution scenario, manufacturers or users can develop extended OPC UA information models to suit their individual requirements. In practical terms, this is often seen in situations where a user needs to integrate multiple sites, each running different devices and interfaces, with a central platform and specific set of applications. An information model designed to accommodate their specific requirements can standardize data access across all of these sites, making IoT application development both simpler and more efficient.
- The configuration effort required for an IoT solution can be reduced further – and operations streamlined – by combining standardized OPC UA information models with suitable technologies and design patterns.
Configurability and mapping
While extended information models offer many benefits, using them efficiently in practice means that OPC UA servers need simple and flexible methods for handling these models. Accordingly, the user needs to be able to load arbitrary information models and map data sources to an address space in the OPC UA server that corresponds to the information model. This mapping functionality is essential for the efficient construction of industrial data spaces in innovative IoT solutions.
Later in this article, we show how Softing implements this functionality in its standard products. First of all, however, we will use three key aspects to investigate how an OPC UA server with mapping functionality can be implemented as part of a larger IOT solution and can then optionally interact with other components. This article will only briefly touch on these aspects because the topic of architecture itself deserves its own, much more detailed discussion.
Mapping near to the data source
In the vast majority of IoT solutions, the task of applying semantics to unstructured data or mapping it to the address space of an OPC UA server needs to be done as near as possible to the data source and not within a centralized data center or cloud. There are many reasons for this, one being applications that already consume data locally at the edge. Another example would be the kind of structured analysis that encompasses multiple levels, such as calculating OEE scores for individual machines, a site, or the entire group of companies. The sheer volume of data produced by devices and automation networks may also create its own problem, with the available budget or bandwidth preventing the transfer of this data to a central platform in its entirety.
Unified namespaces
In the context of IoT solutions, a “unified namespace” is a design pattern for software solutions. With a unified namespace, OT and other data can be provided by standard IT technologies, which ensures the efficient use of this data by application developers and data scientists. In a real-world implementation, this could involve (alongside other components) the use of an MQTT broker with JSON encoding for user data, for example, which in turn accesses an OPC UA server as its data source. For the publication and consolidation of data within a unified namespace, these tasks can be simplified and largely automated by utilizing OPC UA servers based on standard information models.
UA Cloud Library
Once users have decided to deploy OPC UA information models for their IoT solution, model management is the next question that needs to be addressed. Developed jointly by the OPC Foundation and CESMII, the UA Cloud Library is a database for OPC UA information models and address spaces [https://opcfoundation.org/markets-collaboration/cloudlib/]. An HTTP REST interface is provided to simplify the sharing, exploration and distribution of information models, whether manually or with automated tools. Two access models are provided, one being the global UA Cloud Library instance hosted by the OPC Foundation [https://uacloudlibrary.opcfoundation.org/]. Alternatively, users who want to operate their own instance can run the UA Cloud Library as an open-source project [https://github.com/OPCFoundation/UA-CloudLibrary/]. Whichever deployment option is chosen, the UA Cloud Library helps users develop and scale their IoT solutions when facing the challenge of ensuring the efficient management of OPC UA information models and address spaces.
Product portfolio
Launched several years ago, the Secure Integration Server (SIS) from Softing Industrial is a Windows application that acquires data from several sources via OPC UA and aggregates this data within a single OPC UA server. SIS then permits the configurable mapping of the data sources to match loadable OPC UA information models. A second product, the edgeAggregator, was then made available, which offers a comparable range of functionality as a containerized software module.
Recently, Softing once again expanded its portfolio by adding OPC UA mapping functionality to its edgeConnector family and the edgeGate hardware product. Alongside OPC UA and MQTT, these products can source data from a wide range of industrial protocols. Access is provided to Siemens and Allen-Bradley controls, for example, as well as Siemens Sinumerik or Fanuc CNC machines and data sources that “speak” Modbus. In all of these products, data source mapping definitions can be entered manually via a GUI or supplied by an automated system via HTTP REST.
For customers who need to collect machine data and configure its flexible mapping to OPC UA address spaces, Softing’s portfolio therefore ensures they can pick a product that matches their preferred infrastructure and operating model.
Reference project
A large company from the minerals industry launched a broad-based digitalization initiative with a multi-year timescale. As part of this project, the company aimed to develop at least 20 IoT applications and roll these out to more than 100 production sites via a cloud platform (AWS). The company had originally planned to transfer all the unstructured machine data collected to the cloud platform but this quickly proved impractical. Abandoning the idea, the customer realized they needed to deploy local OPC UA information models at machine level to standardize the data before it left the edge. In this scenario, OPC UA information models provide a layer of abstraction between OT and IT, ensuring that all of the sites “look the same” to IoT application developers and data scientists alike. Knowing that the information models deployed would change during the solution lifetime, the customer needed a set of tools that would guarantee efficient, centralized management. This is provided by the UA Cloud Library, which the customer runs themselves on the AWS platform.
Summary
The use of OPC UA information models delivers significant benefits to device manufacturers, application engineers and customers alike. Softing Industrial offers a comprehensive portfolio of connectivity products that utilize a range of industrial protocols to collect device data at the edge and make this data available to client applications via user-configurable OPC UA servers. Softing’s approach here expands the functionality of the critical component of machine connectivity, and establishes a foundation for the flexible and efficient construction of industrial data spaces in innovative IoT solutions.
