Modern industrial enterprises face a critical limitation: legacy integrations between SCADA and MES systems have become a bottleneck, hindering the implementation of AI analytics and effective data management. The main obstacle to digital transformation is data silos on the production floor, where information from sensors and controllers loses context (timestamps, data types, metadata) when transferred to the IT environment, requiring constant manual mapping.
Why the classic SCADA-MES link blocks industrial analytics development
In classic automation architecture, data flows from field devices (sensors, actuators) to PLCs, then to SCADA, and finally to manufacturing execution systems (MES). At each stage, data undergoes transformation, often resulting in the loss of useful context. Proprietary protocols of the past were primarily designed for high-speed value transmission within closed industrial networks, without regard for comprehensive semantic description.
Consider a practical scenario: you attempt to transmit industrial vibration sensor data directly to a cloud analytical system for predictive maintenance. Without a unified semantic layer, data arrives as an array of raw numbers. The cloud system does not know which node the reading belongs to, the sampling frequency, or the units of measurement. Due to the lack of semantic interoperability, the context is lost.
The second problem is architectural vulnerability. Attempting to provide IT systems with direct access to SCADA without adhering to network segmentation requirements creates a serious risk of compromising the operational technology (OT) environment. Connecting the corporate segment to legacy industrial equipment poses a direct threat to production continuity.
OPC UA semantic interoperability: turning raw signals into understandable IT data
The OPC UA (OPC Unified Architecture) standard, developed by the OPC Foundation, is a platform-independent architecture that ensures secure and reliable data exchange between equipment, SCADA/MES, and corporate systems. Its primary value lies in providing semantic interoperability.
OPC UA uses an object-oriented information model. Each data point is not just a memory address, but an object with defined properties and metadata. For example, a sensor in the OPC UA model is described as an object possessing a current value, engineering units, measurement range, data validity status, and a link to a specific technological node.
It is important to understand that OPC UA is not a standalone tool for AI or a magic solution that automatically updates legacy hardware without additional gateways. It is a standardized data normalization layer. It allows wrapping raw streams from heterogeneous PLCs into a unified information model that can be seamlessly consumed by analytical platforms.
Secure integration architecture: IT/OT alignment with ISA/IEC 62443 and NIST SP 800-82
Integrating operational technology (OT) and corporate IT environments requires strict protection policies. Network segmentation is a fundamental control for industrial cybersecurity. The ISA/IEC 62443 standard covers over 20 industries using operational technologies and is key to the cybersecurity of automation systems. It requires network partitioning into zones with different trust levels and the creation of secure transition channels.
According to NIST SP 800-82 guidelines for protecting critical infrastructure, IIoT security requires accounting for the specifics of legacy equipment that cannot always be updated. Direct connections between corporate networks and the production environment are unacceptable. Secure architecture involves deploying intermediate OPC UA gateways that act as buffers in a demilitarized zone (DMZ).
Computing distribution: when to process data at the edge versus the cloud
The reliability of an IoT solution is established at the "device → edge → cloud" architectural stage. Following AWS Well-Architected IoT Lens best practices, it is essential to correctly distribute telemetry processing.
Edge computing is necessary where minimal latency is critical or where the volume of raw data is too large for transmission. A local gateway with OPC UA support can aggregate data locally, filter it, and send only processed aggregates or anomaly alerts to the cloud. Cloud platforms, in turn, are ideal for long-term storage of historical data and analytics.
| Parameter | Edge processing via OPC UA | Cloud processing (Cloud Analytics) |
|---|---|---|
| Critical latency | Minimal (milliseconds) — for local security scenarios and rapid response | High (seconds/minutes) — for long-term trends |
| Data volume and traffic | Filtering, aggregation, and compression of local traffic; transmitting only anomalies | Storing terabytes of historical data for ML model training |
| Security context (ISA/IEC 62443) | Local secure OT environment, minimizing open external ports | Isolated cloud environment with authentication via OPC UA certificates |
Evolutionary transition: from telemetry unification to predictive maintenance
Successful scaling of a device fleet requires a systematic approach. The transition should begin by installing integration gateways in critical areas where data normalization will yield the fastest results.
For the implementation of such complex integration projects, the development of IoT/embedded systems, and architectural consulting, the expertise of Softengi (part of the Intecracy Group alliance) is engaged. Additionally, the alliance portfolio includes the AZIOT Platform — an IoT platform for infrastructure monitoring, edge data collection, and integration with SCADA (supporting MQTT and Modbus protocols).
When the need arises to build a corporate layer for managing normalized data, auditing, or integration processes, solutions are built on the UnityBase platform. This is a full-stack JavaScript low-code platform, which is a joint development of the companies within the Intecracy Group alliance (where InBase acts as a key, but not the only, developer). UnityBase provides a unified Domain metadata model that combines data generated by REST API and strict access control policies (RBAC/RLS). For high-load systems or projects with heightened security requirements, the official page recommends the Enterprise or Defence editions of the UnityBase platform.
Such architecture allows businesses to transform smoothly: from local reading collection to centralized analytics, the creation of digital twins, and predictive maintenance based on clean, semantically understandable data.
FAQ
How does OPC UA solve the compatibility problem between equipment from different vendors (Siemens, Beckhoff, Rockwell)?
OPC UA provides a platform-independent architecture and semantic interoperability. Instead of translating raw memory addresses of different PLCs, the protocol allows data to be normalized into a unified object-oriented information model understandable to IT systems, preserving all metadata and engineering units.
Does implementing OPC UA replace the need for network segmentation under the ISA/IEC 62443 standard?
No, it does not. According to ISA/IEC 62443 standards and NIST SP 800-82 recommendations, IT/OT segmentation is a fundamental cybersecurity control. The implementation of OPC UA must be accompanied by the creation of zones, a demilitarized zone (DMZ), and the configuration of firewalls.
How can data transmission from OPC UA to cloud platforms like AWS or Azure be organized without losing metadata?
According to IoT solution architecture (e.g., AWS IoT Lens), processing is distributed between the Edge and the Cloud. Intermediate Edge gateways read the OPC UA object model locally, filter the data, and securely transmit only the necessary information to the cloud with the context (metadata) preserved.