Software-defined automation ushers in IT-like engineering

Industrial automation, historically driven by reliability and efficiency, has successfully automated production worldwide. However, evolving consumer demands for greater customization are now straining existing device-centric automation and control systems. This escalating need for adaptability pushes production complexity to its limits.

To overcome these challenges, we must redefine production processes through a software-first approach.

Software-defined approaches thrive in industries where constant change is the norm, offering solutions through methods like developer operations (DevOps) for rapid iteration, optimization and cross-location collaboration.

While stability is paramount in production, adopting software-defined automation (SDA) integrates decades of automation expertise with pioneering IT tools and methods to scale industry-wide. This isn't about a single solution, but a paradigm shift: a mindset and a set of shared principles that bring adaptability and scalability to production by converging information technology (IT) and operational technology (OT).

SDA's core objective is to enhance flexibility and adaptability by elevating abstraction and automating both runtime and engineering processes. At runtime, this involves decoupling automation applications from embedded hardware, running them in containerized environments alongside data and AI applications. With central management and versioning on a standardized software infrastructure, scalability is limited only by compute resources.

For engineering, SDA champions modular, reusable and standardized software components, tools and modern development practices. This empowers engineers to tailor their toolchains, modularize application programs and even deploy them remotely. Rapidly deployable virtualized controllers and reusable engineering modules unlock true adaptability and flexibility, even in highly complex production settings.

The virtual controller market, a key SDA component, could reach $4.5 billion by 2035, according to ABI Research's Industrial Automation: Software-Defined Automation (SDA) market data report. This projection underscores SDA's imminent widespread adoption, particularly where modernization and flexibility are crucial.

Same runtime environment, different structure

The programmable logic controller (PLC) revolutionized industrial automation, enabling engineers to "rewire" plants in software rather than physically altering control circuits. PLCs, powered by increasingly powerful hardware, facilitated advanced control algorithms and have been a production floor staple for decades. While many companies augment production with industrial PCs and sensors for machine health and quality tracking, significant challenges arise when businesses seek to scale or fine-tune production.

OT can learn from the world of IT with virtualization. By moving the runtime stack of a PLC into a containerized, or virtual, system the application runtime can be decoupled from embedded hardware. For example, virtualized PLCs (vPLCs) function the same way as a regular PLC but have the possibility to run on standardized software infrastructure that can be utilized on a greater variance of compute resources. This also makes it easier to communicate between applications, maintain version control across your factory floor and manage them centrally. Virtual PLCs enable a convergence of the control and other skills on the production floor.

A virtualized controls approach offers significant long-term advantages. With runtime in a digital environment, automation experts can rapidly deploy new instances and programs without waiting for physical hardware. Virtual PLCs introduce data-driven decision-making and modern architectural setups to the automation layer, mirroring decades of software development practices. Integrated with simulation tools, new automation schemes can be defined, tested and monitored post-rollout. Real-world data feedback helps validate simulated changes, bridging the real and digital worlds. Containerization enables rapid, low-risk changes alongside running operations, with easy system rollbacks if errors occur. Program revisions can be deployed and monitored effortlessly, validating against simulations—an ideal IT-like engineering approach.

DevOps for industrial automation

As factories adopt data-center methodologies for control via virtualization, automation engineers will leverage software development practices to enhance flexibility and scalability. Traditionally, automation logic development adhered to electrical engineering standards, which made sense when control logic abstracted electrical processes within a PLC. A key shift from modern software development is adopting document-based, human-readable, plain-text source code for automation logic, granting engineers significantly more programming flexibility.

This shift in the approach to application code and programming languages is just one facet in the broader paradigm shift. IT-like engineering is about changing how we approach the development of automation in general. Businesses should not think of programming as a one-time thing that will be left undocumented, untouched and unoptimized until an issue arises. Instead, as with software development, the code needs to be maintained, adapted and optimized over time.

There are two key developments to make this shift possible: new tools combined with industrial automation and a mindset change. This can be summarized with the concept of DevOps for industrial automation. Software development tools can help adaptability during both engineering and operations through standardization because, the more easily code can be written, versioned, shared and collaborated on, the more easily it can be optimized. This aids the shifting mindset.

Getting the logistics of development out of the way of engineers lets users think longer-term about what they are implementing. One way may be fast in the short-term, but adding effort up-front could save headaches down the road. Furthermore, employing these new technologies increases the potential pool of qualified talent, as they are provided with a familiar development environment and can focus on understanding the physics and mechanical aspects of automation engineering.

Get your subscription to Control Design’s daily newsletter.

While the goal of machine operation remains constant, the methodology evolves. Adopting a software developer's approach unlocks continuous benefits from existing automation engineering practices. Object-oriented programming (OOP), for instance, significantly improves project standardization and modularization; a motion routine for a base valve can be adapted via inheritance without altering its original structure.

Unit testing provides rapid validation of application logic without requiring a physical PLC. Furthermore, a text-based, API-first strategy with native Git integration facilitates constant sharing of programming successes among automation engineers. This collaborative sharing of work and best practices extends beyond the factory, inspiring novel approaches that individual "solution heroes" might overlook.

Connected, but decoupled

These two prongs of SDA do not need to be done simultaneously. This can be a lot to do at once. The good news is it doesn’t have to be. Some companies might already be adopting vPLCs and could implement more IT-like engineering afterwards. Others might be deploying IT-like engineering and gradually shift into vPLCs as they expand operations. Being able to work on one while the other hums along de-risks optimizations and reduces the downtime for changes, especially also from a mindset perspective.

Companies are successfully implementing SDA in production. In the automotive industry, for instance, SDA simplifies production scaling and the convergence of control, data and AI applications. It also facilitates growth: instead of investing in every potential hardware piece, companies can gradually expand production capacity, ensuring new hardware integrates seamlessly into existing lines.

Some manufacturing facilities are also seeing the value of hiring a new type of developer—the software automation engineer—bringing in new skills and approaches like the reuse of open source or the use of artificial intelligence. A new ecosystem of automation engineering is emerging and providing more pathways to success beyond an individual solution hero or factory. Some are embracing the collaborative mindset of software development through open-source projects while others are reaping the benefits of pipelines and more efficient testing to optimize entire control setups. There are so many possibilities that are emerging from the shift to SDA.

Conclusion

Software-defined automation marks a paradigm shift for the industrial world. A software-defined methodology enables a new level of agility and flexibility. SDA becomes the connective layer between digital twins, AI decision models and the physical control level. It makes use of runtime and engineering as strategic levers to create more adaptive factories.

The journey toward SDA promises a wide spectrum of benefits, some immediate, others evolving as new mindsets and system architectures mature. Products are not manufactured in a vacuum, and markets shift rapidly. By embracing software-defined principles and fostering collaboration, businesses can significantly enhance their competitive advantage.

Ultimately, a product design or recipe entering the production environment could initiate and automatically adjust the process on a shop floor. That might be a long way away, but SDA lays the groundwork for that and many other potential futures. It fosters flexibility from the core.