Since the inception of GPT and AI chat models, the tech space has become somewhat of a impersonation of the movie Everything Everywhere All at Once. It has boosted the development of so many products and projects that the inter-connection is becoming a bottleneck.
With various protocols coming in the links the skills to agents to services, we need something the connects various edge data to each other while being independent of the vendor. The need for an open source aggregator has become imminent. Now this missing bridge is OPC-UA. It quite literally works like a connecting various edge devices, from robots to ESP32 based weather stations and everything. It makes the data transmission and communication between endpoints seamless and hassle free.
The legacy protocols like Modbus held quite well back in the days but today they need an upgrade. From AI models to cloud based SCADA systems all use structured and typed data which is a bit tricky implementing over modbus. The traditional system demands for a pool of individual protocol converters to be installed to handle inter unit data exchange. This calls for introduction of OPC-UA to the traditional systems.
It is designed not just as a bridge but a central processing unit. The traversing data can be filtered, processed or harmonization can be done on the fly. The gateway is engineered for maximum industrial versatility, offering dual Modbus support that enables simultaneous connection management for both legacy serial (RTU) and modern Ethernet (TCP) devices. At its core, it features an embedded OPC UA server powered by open62541 that resides directly on the edge device. The future expansions add support for various other protocols such as CAN. Which brings the protocol conversion from N:N to 1:1.
The primary goal of the aggregator is to simplify the infrastructure stack. Instead of an AI model having to query ten different Modbus slaves across different serial ports and IP addresses, it queries one OPC UA server. The gateway utilizes a JSON-based configuration system. This allows system integrators to define the entire industrial environment in a single file. By editing config/config.json, users can map specific Modbus registers to human-readable OPC UA NodeIDs.
Example Configuration Snippet:
JSON
{
"modbus": {
"type": "rtu",
"address": "/dev/ttyUSB0",
"slave_id": 1
},
"opcua": {
"port": 4841
},
"data": [
{
"mb_register": 40001,
"mb_dlen": 2,
"ua_nodeid": "ns=1;s=Motor_Temperature",
"ua_dtype": "float"
}
]
}
This architecture ensures that the AI developer never needs to know the Modbus register address 40001. They simply subscribe to Motor_Temperature. This abstraction layer is critical for scaling AI deployments across multiple factories with varying hardware versions.
Now the added benefit of the OPC-UA gateway is data filtering. As data is the fuel for AI, but "raw" industrial data is often noisy and redundant. Sending every single Modbus poll (which may occur every 100ms) to a cloud-based AI model is expensive and inefficient.By utilizing the gateway's Polling Architecture, we can implement edge side logic. The gateway operates on a configurable poll_interval_ms. Within this loop, the aggregator can perform:
1) Only update the OPC UA node if the value has changed beyond a certain threshold.
2) Convert raw values into SI units
3) Use simple logic to flag "out-of-range" data
This can help solve the data logistics problems for the AI and cloud infra today. This will help fast track and optimize the operations on the data by a good margin. OPC-UA is not just another industrial protocol. It is a practical way to simplify communication between modern AI systems and traditional industrial hardware. The aggregator creates a single standardized interface for the entire system eliminating the need of building separate integrations for every protocol. This reduces infrastructure complexity, improves scalability, and makes industrial data easier to process and use. OPC-UA can serve as the common layer, aiding the industries in adopting AI, cloud systems, and edge computing.
