Modern industrial enterprises are actively integrating IT and OT (Operational Technology) environments. This process requires a standardized approach to data architecture, security, and scaling solutions in the Industrial Internet of Things (IIoT) sphere. However, in practice, CTOs and engineers face a significant challenge: how to integrate legacy equipment operating on closed proprietary protocols with modern analytical systems. Secure data transmission and visualization at the decision-making level often become a bottleneck for the entire project.
The conflict between closed legacy machine protocols and the open standards of modern platforms creates a gap in information transfer. Attempts to connect critical equipment directly to IT networks without proper segmentation create security risks. To build a reliable system, it is necessary to design every stage of signal transmission: from physical contact on the production line to the final graph on a monitor screen.
Anatomy of IIoT architecture: from physical sensor to cloud storage
Industrial IoT is not just a set of "smart" devices; it is a multi-level technology stack. According to the AWS Well-Architected IoT Lens recommendations, the reliability of any IoT solution is established directly at the architecture design stage, which covers the entire data path: from the end physical device through the network edge (Edge) to cloud or on-premises storage. An error at the lower level will result in all subsequent analytics losing meaning due to the "garbage in, garbage out" phenomenon.
A typical path for industrial data consists of the following steps:
- Physical level: Temperature, pressure, vibration sensors, or flow meters generate analog or digital signals.
- Controller level: Programmable Logic Controllers (PLC) collect these signals for direct process control in real-time.
- Edge level: Industrial gateways read data from PLCs, perform initial filtering, and convert protocols.
- Data transmission level: Message brokers ensure the delivery of normalized telemetry.
- Platform level: Cloud or on-premises systems where data is accumulated, analyzed, and stored in Time Series DBs.
The choice of communication technology at each stage depends on the business case. There is no universal standard: in some conditions, classic Ethernet is optimal, while in others, wireless solutions are better. Deploying 5G technology should be considered primarily where ultra-low latency is required for autonomous mobile robots or critical control.
The legacy equipment problem and the role of OPC UA in industrial data unification
A significant obstacle for many enterprises is machinery and controllers installed decades ago. They perform their mechanical functions but often operate via outdated interfaces and protocols that lack built-in security or encryption mechanisms, and data is transmitted as "raw" registers without semantic description.
To solve this problem, data normalization from various sensors is applied before transmission to an analytical system. The main tool for unification has become the OPC UA (OPC Unified Architecture) standard. According to the OPC Foundation, it provides a platform-independent architecture for secure data exchange between industrial equipment and corporate systems (SCADA/MES).
OPC UA transmits not just a numerical value, but extended metadata: value type, units of measurement, timestamp, and signal validity status. This allows transforming a disparate stream into a structured information model that modern IT systems can easily process.
Edge computing vs. Cloud: where to process telemetry without losing performance
One of the key decisions during design is choosing between local data processing on an Edge device and sending it to the cloud. Transmitting all raw signals directly to the cloud can quickly overload communication channels and lead to unjustified infrastructure costs.
For example, a vibration sensor taking thousands of measurements per second generates a massive data array. Instead of transmitting it entirely, primary analysis is performed on the Edge gateway, and only an aggregated indicator with a set periodicity is sent to the cloud. At the same time, if a local algorithm detects a critical spike, the system reacts instantly.
Cloud resources remain indispensable for storing multi-year historical data, identifying long-term trends, and building predictive maintenance algorithms. However, scaling a fleet of IoT devices requires a systematic approach to automated provisioning, firmware updates, and equipment status monitoring.
Security of OT environments: network segmentation and critical infrastructure protection
When integrating IT and OT, a fundamental conflict of priorities arises. According to the NIST SP 800-82 Guide to OT Security, in the field of operational technology (OT), system availability is often prioritized over data confidentiality. Stopping a production process due to a false positive from a security system can lead to critical consequences.
Protection is complicated by the presence of legacy equipment that does not support modern cybersecurity methods. Accordingly, security is built at the network architecture level. The main tool is network segmentation to separate critical production equipment from the corporate IT network. A reliable foundation for managing these risks is the ISA/IEC 62443 series of standards, which defines zones, communication channels, and cybersecurity requirements for industrial automation systems, and is applied in over 20 industries.
Designing a dashboard: how to turn raw signals into actionable insights for business
The ultimate goal of an IIoT project is to provide the decision-maker with clear and relevant information. Data that has traveled from the sensor to the platform must be visualized on a dashboard. It should not be overloaded with charts; its task is to reflect the real state of processes through key indicators (equipment efficiency, downtime, energy consumption, maintenance forecasts).
To build such complex enterprise-level solutions, specialized contractors with deep development expertise are engaged. In particular, Softengi (a member of the Intecracy Group technology alliance) provides development of custom IoT/embedded solutions, their integration with SCADA systems, and cloud architecture design, helping enterprise clients reliably connect the physical sensor layer with analytical panels.
Maturity levels of enterprise IIoT infrastructure
| Maturity level | Description and infrastructure characteristics |
|---|---|
| Level 0 (Chaotic) | Data is collected manually or isolated within local SCADA systems of individual workshops; legacy equipment is not connected to the general network. |
| Level 1 (Basic monitoring) | Basic telemetry is available; data is collected in a centralized database, but there is no normalization or unified transmission standard. |
| Level 2 (Standardized) | OPC UA is used for data unification; Edge devices are implemented for initial filtering; IT and OT networks are segmented. |
| Level 3 (Integrated) | Data is visualized on a dashboard in real-time; automatic device provisioning is configured; integration with MES/ERP is established. |
| Level 4 (Predictive) | Predictive maintenance scenarios and digital twins are implemented based on long-term analysis of historical data. |
Building a reliable IIoT architecture is an evolutionary process. Starting with segmentation and standardization of data exchange via OPC UA, an enterprise lays the foundation for transitioning to real-time analytics and predictive production management.
FAQ
How to connect old industrial equipment to a modern IIoT platform without replacing controllers?
Industrial Edge gateways are used for this. They physically connect to legacy controller interfaces, read data via local protocols (e.g., Modbus), and convert them into modern standardized formats, such as OPC UA, for subsequent transmission to the IT network.
Why can't all sensor data be sent directly to the cloud, and how does Edge computing help with this?
Sending the full stream of raw data directly creates excessive load on communication channels and increases storage costs. Edge computing allows for primary processing, aggregation, and filtering of telemetry directly at the source, sending only useful information or critical alerts to the cloud.
What security standards are fundamental for protecting data when integrating IT and OT networks?
The main guidelines are the NIST SP 800-82 guide and the ISA/IEC 62443 series of standards. They define the principles for protecting industrial systems, among which strict network segmentation—separating the operational network from the corporate one—plays a key role.