Modern industrial enterprises face the challenge of integrating disparate data, where OPC UA serves not just as a communication protocol, but as a critical foundation for scalable analytics and predictive maintenance. Classic SCADA-MES integrations create data silos that complicate the implementation of modern analytical tools and artificial intelligence. Without semantic data compatibility, engineers must dedicate resources to manual mapping of thousands of tags, which lose context as they travel from physical sensors to the corporate IT network.
Why the classic SCADA-MES automation pyramid limits analytics scaling
Traditional industrial automation architecture rigidly separates levels, from field devices to supervisory control (SCADA) and manufacturing execution systems (MES). In the industrial IoT era, this hierarchy often becomes a barrier. Heterogeneous programmable logic controllers (PLC) use proprietary closed protocols, forcing engineers to build complex point-to-point integrations.
The core issue is that classic protocols transmit raw values without descriptions. When data reaches SCADA or IT systems, it loses critical metadata: units of measurement, threshold values, calibration info, and precise source timestamps. Without a unified semantic description, data loses context when transmitted to the cloud, making the deployment of predictive maintenance models or computer vision systems (such as Ionbond AI Visual Inspection by Softengi for quality control) a resource-intensive process of individual configuration for every node.
Semantic compatibility of OPC UA: converting raw tags into structured data
The OPC Foundation developed the OPC Unified Architecture (OPC UA) standard as a platform-independent architecture for secure and reliable interoperability in industrial systems. OPC UA operates on any operating system (Linux, Windows, RTOS) and ensures full semantic data compatibility.
Normalizing data streams from heterogeneous PLCs from various manufacturers into a single format allows for a transition from flat tag tables to object-oriented information models. Instead of an abstract tag like DB10.DBD20, the IT system receives a structured object Pump01.Temperature with built-in attributes: current value, unit of measurement, validity status, and a timestamp generated by the controller. This allows analytical platforms to automatically interpret telemetry.
Architectural balance: handling telemetry between edge and cloud
According to AWS Well-Architected IoT Lens practices, the reliability of an IoT solution is established at the device-edge-cloud architecture stage, where it is crucial to correctly distribute telemetry processing. Moving critical signal processing to edge devices reduces latency and ensures production autonomy during cloud connectivity loss.
At the edge level, primary filtering of high-frequency data, local execution of AI models for anomaly detection, and information buffering are performed. Long-term analytics and global planning remain in the cloud environment. To achieve this architectural balance, the AZIOT Platform (a product of AZIOT) can be used, which ensures data collection at the edge via industrial protocols, local processing, and secure transmission to corporate IT systems.
Cybersecurity of the OT perimeter: OPC UA within ISA/IEC 62443 and NIST SP 800-82 standards
Integrating industrial data (OT) with the corporate IT segment creates significant risks for technological processes. According to the NIST SP 800-82 Guide to OT Security, in the field of operational technology (OT/ICS/SCADA) security, system availability is often prioritized over data confidentiality. Stopping a production line due to blocked legitimate traffic or a cyberattack causes catastrophic losses.
The ISA/IEC 62443 series of standards covers over 20 industries where operational technologies are used. Implementing network segmentation between the shop floor equipment level and the corporate IT network according to this standard is mandatory. OPC UA supports encryption, authentication, and access control, but it does not solve all security issues without proper configuration of firewalls, demilitarized zones (DMZ), and edge gateways. Legacy equipment that does not support modern protocols does not require immediate replacement; it is isolated in separate subnets using industrial gateways.
Migration strategy: preparing infrastructure for digital twins
Transitioning to the digital twins concept requires a consistent increase in infrastructure maturity. The following scale helps assess the state of industrial data integration:
| Maturity level | Integration status description | Technological approach |
|---|---|---|
| Level 0 (Chaotic) | Data is locked within local PLC/SCADA. No integration, calendar-based maintenance. | Proprietary equipment manufacturer protocols. |
| Level 1 (Point-to-point) | Direct SCADA to MES connection via proprietary drivers. Constant manual tag mapping. | Specialized access drivers, flat tags. |
| Level 2 (Standardized) | OPC UA implemented at the edge level. Data is normalized, but analytics are limited to the local loop. | OPC UA servers, local edge gateways. |
| Level 3 (Convergent) | Segmented IT/OT network per ISA/IEC 62443. OPC UA data transmitted to the cloud with semantic preservation. | OPC UA, MQTT, DMZ routing. |
| Level 4 (Intelligent) | Digital twins and predictive maintenance built on unified data streams. | AI/ML models, enterprise integration platforms. |
Building the upper IT layer that receives and processes normalized data for business users requires a reliable technological foundation. An example of such a foundation is the use of the UnityBase platform. Thanks to the Domain metadata concept, the platform allows for the creation of a managed enterprise domain model and the generation of REST API for integration with IIoT platforms. For high-load critical infrastructure objects, deploying Enterprise (EE) or Defence (DE) editions of the UnityBase platform is appropriate, ensuring strict access control (RBAC/RLS) and detailed audit of actions in a secure on-premises environment.
FAQ
How does OPC UA solve the compatibility problem between PLCs of different brands (Siemens, Beckhoff, Rockwell)?
OPC UA acts as a platform-independent translator. Instead of writing separate drivers for each brand, an OPC UA server is installed at the edge level (or native capabilities of modern controllers are used). It normalizes proprietary protocols into a unified object model with a common semantic description.
How to ensure security when transmitting data from OPC UA to the cloud according to ISA/IEC 62443?
Security is ensured through strict network segmentation (Zones and Conduits). Data from the OT perimeter should not go to the cloud directly. An intermediate edge gateway in a demilitarized zone (DMZ) is used, which collects data via OPC UA, encrypts it, and transmits it externally through a secure outbound channel.
Is it necessary to replace old controllers (legacy) that do not support OPC UA natively?
No, replacing functional legacy equipment is not mandatory. Data from old controllers is collected via their native protocols (e.g., Modbus) using local industrial edge gateways, which convert this data into an OPC UA information model.