IIoT technologies are addressing a space that is already turned-on to the benefits of machine-to-machine communication, remote access and intensive data gathering to feed statistical process control, manufacturing management and continuous improvement initiatives. In fact, the data flows throughout modern industrial enterprises are already organized into a mature hierarchy.
At the supervisory control and data acquisition (SCADA) layer and below, industrial Ethernet protocols have evolved to meet specific demands associated with process automation. One particular requirement is support for deterministic control, often in real-time with only sub-second latencies permitted. Industrial Ethernet protocols have also been adopted to satisfy the growing appetite for data among today’s industrial enterprises. These are helping them to meet customers’ demands more quickly, with increased quality and lower costs.
Such commercial goals are well established. However, historically, industrial connectivity has tended towards reliance on custom solutions, based on proprietary standards, that are expensive to implement, complicated to maintain and quite difficult to scale. Although IIoT technologies could help to change all this, the industrial automation scene is already trying to overcome its known limitations. Industrial Ethernet protocols have coalesced around a few popular specifications - such as Profinet, for example.
Business-Critical Industrial Ethernet
Profinet is well tuned to industrial automation tasks - while supporting growing demand for richer machine-to-machine communications and data capture, it is also adaptable to meet specific needs of industrial processes. A large set of application profiles ensure families of devices can use Profinet in the same way and thereby rely on interoperability and interchangeability of devices and systems from various different manufacturers. Some examples of this include PROFIdrive for drives/motors, PROFIenergy which is applicable to energy management, and PROFIsafe for communication relating to functional safety. Moreover, the Profinet I/O profile provides a means of connecting non-smart devices like mechanical switches to monitor status and send commands.
Profinet is also designed to handle special industrial requirements, such as safe operation in hazardous environments where Ethernet cannot be used. Profinet handles this through proxies that allow safe fieldbus terminals to connect equipment in hazardous areas to the Ethernet-based infrastructure. A wide variety of fieldbus specifications are supported in this way, including PROFIBUS DP, PROFIBUS PA, Interbus, DeviceNet, Foundation Fieldbus, etc., plus non-smart devices using profiles such as Profinet I/O.
The success of Profinet, with upwards of 17 million nodes already installed, suggests that the supervisory control aspect of connections between SCADA systems and process automation equipment is well catered for and working effectively. IIoT technologies could help enrich the other side of the equation - namely data acquisition - and, in particular, making data from the manufacturing activities relevant and accessible throughout the enterprise IT infrastructure above the SCADA layer.
What can IIoT Bring to the Table?
From a commercial perspective, IIoT technologies will interest enterprises as a means of delivering digital transformation - ultimately to increase manufacturing productivity, improve business efficiency, and drive innovation.
Object linking and embedding for process control unified architecture (known as OPC UA) is an emerging Ethernet-based protocol designed to deal with communication of industrial process data, alarms, events and historical data. It is platform neutral, with built-in security, and capable of supporting autonomous and intelligent systems. Through these properties, it provides a platform for open and vendor-independent communication from the controller level to link programmable logic controller (PLC) and sensor data with SCADA systems, MES, ERP, and the Cloud.
Lightweight Protocols, Open Standards
Message queuing telemetry transport (MQTT) could potentially become a pervasive protocol, unifying the operational technology (OT) and information technology (IT) domains. Originally MQTT was designed as a lightweight, data-agnostic, low-overhead protocol for low-bandwidth data links in the oil and gas industry. It is now used by messaging applications like Facebook Messenger, as well as the Amazon and IBM IoT platforms.
MQTT delivers the scalability and flexibility expected of IIoT technologies, by operating a publish/subscribe model that decouples data producers from data consumers. This is fundamentally different to the traditional M2M principles employed in the industrial domain, which are typically master/slave interactions, or polled responses, where data is repeatedly requested from specific sources. In this traditional model, introducing new data sources such as sensors, or changing or extending applications to work with larger data sets, is a complex and exacting challenge.
In the publish/subscribe approach used by MQTT, a data producer (such as a sensor) simply publishes its data as soon as it is ready. That data is then generally available to any subscribing consumers. New equipment can be introduced or subscriptions changed with minimal intervention. Publishing data on a report by exception (RBE) basis liberates network bandwidth for use by critical OT equipment and enterprise systems. Furthermore, the low packet overhead and straightforward functionality contribute to its simplicity - with the header constituting a meagre 2bytes, payload being application specific, and the only actions supported being connect, disconnect, publish, subscribe and ping.
A number of other publish/subscribe protocols could prove useful for data acquisition in IIoT scenarios. Data distribution service (DDS) is an open protocol that requires no centralized broker, like an MQTT server, and so permits simpler network management as well as faster communication. Sub-millisecond resolution makes DDS suitable for real-time and embedded systems, plus business-critical applications.
Among other publish/subscribe protocols, advanced message queuing protocol (AMQP) is already established in the financial services sector and has mechanisms to guarantee completion of transactions. AMQP is not, however, a lightweight protocol, which may hamper its adoption for bandwidth-constrained industrial communications.
Industrial automation, oriented towards SCADA, has been adopting Ethernet-based communications for a number of years, to take advantage of faster speeds, greater efficiency and future-proofing. Convergence on popular specifications (like Profinet) leverages the robustness of communication that is optimized for industrial automation and can handle real-time and mission-critical interactions, as well as ensuring interoperability between modules from different manufacturers that fulfil highly specialized roles.
Internet-based protocols that are suitably rigorous and high-performing may emerge in the future. For now, the key role open to current IIoT technologies is to link the OT and IT domains. This can be achieved by enabling more efficient and lightweight data exchanges, and simplifying access for IT systems to data that is relevant. Scalability and flexibility, which are fundamental principles of IoT implementation, are also important to allow evolution as companies learn how to take advantage of IIoT in their enterprises, and seek additional business advantages.
Rip-and-replace is rarely favoured as an approach in industry, where all investments are considered carefully and expected to deliver a calculated return. Internet-based protocols will enter the industrial automation arena as and when they can deliver a calculable gain. As such, protocols like MQTT, OPC UA, CoAP and others will augment, rather than instantly replace, incumbent protocols like Profinet.