Industrial companies want to manage their operational technology (OT) with the same flexibility they bring to IT. Control system vendors are making that vision more tangible: containerized environments, high-level languages instead of proprietary programming tools, and open platforms for orchestrating machine skills. But how far can the fully software-defined control cabinet really go?
A Two-Layer Architecture
Modern setups split responsibilities across two layers that work independently but in close coordination. At the upper edge layer, a hardened Linux operating system serves as the flexible brain of the plant. Inside isolated Docker containers, high-level languages like Python, Go, or Node-RED handle data-intensive tasks: AI-powered image inspection, dynamic orchestration of machine skills, and data aggregation for MES and ERP systems over MQTT.
Why the PLC Stays
Despite all the flexibility at the edge layer, the programmable logic controller (PLC) has not been displaced from the control cabinet. The reason comes down to three requirements that software-defined environments still struggle to meet reliably: real-time response, functional safety, and high availability. A PLC responds to control signals in a deterministic way, regardless of system load or network conditions. In many industrial processes, that is not optional but a hard requirement.
The result in practice is a hybrid picture: data-intensive, flexible, and AI-driven tasks move to the edge layer. Cyclic control tasks, safety functions, and time-critical processes stay on the PLC. Getting that split right lets you benefit from the strengths of both worlds without compromising plant reliability.