Modern industrial automation systems work with networked real-time data. These connected Industrial Internet of Things (IIoT) systems rely on interoperability so that devices and independent software applications from different vendors can work together seamlessly. They also require performance and flexibility to cope with demanding environments and use cases.

To meet these requirements, the industry needs a clear concept based on proven standards. By combining the best features of the OPC UA, DDS and TSN standards, factory floors can be connected to management floors, sensors to the cloud and real-time devices to manufacturing cells - all based on standardized, interoperable and secure technologies. But how can they work together?

Predominant technology: OPC UA

In industrial automation, OPC UA is the predominant technology for transferring data and information from field devices to the cloud. OPC UA is essentially about creating a distributed object model that end users can use to integrate devices into functioning automation systems. Together with the associated specifications, the standard provides an environment in which devices can be understood and networked. In addition, the OPC UA 'Information Model' (IM) enables self-checking and control of the entire system.

OPC UA's powerful client-server model allows applications - usually clients - to interact with a complex system (implemented as a series of servers). In IIoT, however, the client-server model is not particularly well suited to the information flows required for real-time control and peer-to-peer communication. Currently, these real-time information flows use different fieldbus protocols that do not work together.

In 2018, Part 14 of the OPC UA specification was published, which contains the implementation of the new publish-subscribe extensions. These are used to connect to systems via AMQP, MQTT, Ethernet and UDP, whereby the combination with the new Ethernet standard TSN is possible. The result is a technology that is similar in its characteristics to traditional fieldbuses.

Data Distribution Service (DDS)

The DDS standard developed by the Object Management Group (OMG) for a connectivity framework is used in numerous industries. DDS basically creates an integration framework for software applications and is used in applications that combine intelligent software with fast sensors and actuators. In these systems, the standard serves as a control bus as well as a connectivity framework between the edge and the cloud.

DDS provides a scalable publish-subscribe data bus that provides a protected global data space with a standardized data model for all devices and applications. The data bus and the standardized data model decouple the applications from each other, as they interact less with each other and more with the data bus. They publish or subscribe to topics and read or write data object instances. Security is taken care of at the data level. Applications require permission to join the domain and rely on finely graded authorizations to gain access to a topic or data object.

The standard also supports service-oriented, RPC-based architectures and provides the tools required to define, discover and invoke remote services.

This makes DDS ideal for peer-to-peer communication and real-time information flows. In contrast to other publish-subscribe technologies such as MQTT or AMQP, the implementation here is carried out without a broker or server. Comprehensive QoS (Quality of Service) control makes it possible to adapt each information flow to the respective requirements. The functionality and security properties of DDS are also transport-independent, meaning that it can be implemented across WANs, multicast-capable LANs and TSN networks. In order to provide the required real-time determinism, the transport layer must play its part, which is why the use of DDS with TSN makes sense as soon as critical real-time loops are involved.

The requirements of IIoT systems

Complex systems need both OPC UA and DDS to manage the components and the flow of devices, real-time critical loops, subsystems with predominantly static components as well as the interactions that are more ad-hoc and dynamic. This includes, for example, the static components in a robot during deployment. However, if the robots are mobile, the interactions between the individual robots and the robot groups as well as external agents may change dynamically.

In such a case, one of the aforementioned standards would normally be added to fill the gaps. However, this approach has its limits. The current concept of adding publish-subscribe extensions to OPC UA to enhance the object-oriented client/server structure of OPC UA with retrofitted publish-subscribe concepts cannot provide the required performance or scalability. Similarly, extending the 'decoupled' data-centric model of the DDS data bus with the pub/sub-topics and services for a more tightly coupled object-oriented concept cannot achieve the rich feature set that OPC UA currently offers.

OPC UA and DDS in one architecture

The best features of OPC UA and DDS can be combined in a new integrated architecture. In fact, OPC UA and DDS already coexist and interoperate. The OMG OPC UA/DDS gateway specification provides a mapping of information models that enables pure OPC UA applications to exchange information with pure DDS applications. This gateway specification can now be supplemented by a more comprehensive integration concept.

OPC UA offers interoperability at device level, a well-developed information model (IM) and an extensive ecosystem in industrial automation and requires a proven, high-performance publish-subscribe technology. DDS has extensive data modeling capabilities as well as multiple communication patterns, device-to-device control, transparent scalability and physical layer independence and is widely used in many IIoT industries.