Inferensys

Glossary

OPC UA Pub/Sub

An extension of the OPC Unified Architecture that enables scalable, broker-less, one-to-many or many-to-many data distribution using a publish-subscribe pattern, critical for high-throughput sensor data on the factory floor.
Data scientist building training data pipeline on laptop, data preprocessing visible, technical workspace.

What is OPC UA Pub/Sub?

OPC UA Pub/Sub is an extension of the OPC Unified Architecture that enables scalable, broker-less, one-to-many or many-to-many data distribution using a publish-subscribe pattern, critical for high-throughput sensor data on the factory floor.

OPC UA Pub/Sub decouples data producers from consumers, allowing a single publisher to transmit messages to multiple subscribers simultaneously without a central broker. This contrasts with the client-server model, using User Datagram Protocol (UDP) multicast or broker-based protocols like AMQP and MQTT to achieve the deterministic, low-latency communication required for high-speed industrial automation.

By mapping OPC UA's information model onto efficient transport protocols, Pub/Sub enables the direct integration of field-level devices and sensors with cloud analytics and Manufacturing Execution Systems (MES). This architecture is foundational for Industrial Internet of Things (IIoT) scalability, supporting time-sensitive networking (TSN) for hard real-time control and event-driven closed-loop manufacturing optimization.

SCALABLE DATA DISTRIBUTION

Key Features of OPC UA Pub/Sub

OPC UA Pub/Sub extends the client-server model with a publish-subscribe pattern, enabling efficient, broker-less, one-to-many communication for high-throughput industrial IoT data.

01

Decoupled Publisher-Subscriber Model

Unlike the classic client-server model requiring direct, persistent connections, Pub/Sub decouples data producers (Publishers) from data consumers (Subscribers). Publishers send data to a logical topic without knowing the recipients. Subscribers express interest in a topic and receive data asynchronously. This loose coupling allows systems to scale independently, adding or removing sensors and applications without reconfiguring the entire network. It eliminates the point-to-point spaghetti architecture that plagues traditional SCADA integrations.

02

Transport Protocol Flexibility

OPC UA Pub/Sub is transport-agnostic, allowing architects to choose the best protocol for their infrastructure without changing the application logic. Two primary transports are defined:

  • UADP (UA Datagram Protocol): A lightweight, binary protocol optimized for Layer 2 multicast on time-sensitive networks (TSN). Ideal for high-speed, controller-to-controller communication on the factory floor.
  • MQTT (Message Queuing Telemetry Transport): A TCP-based protocol designed for cloud connectivity and constrained networks. It uses a central broker for store-and-forward messaging, enabling secure data egress to AWS IoT Core or Azure IoT Hub. This flexibility bridges the gap between deterministic shop-floor control and enterprise analytics.
03

Integrated Security Model

Security is not an afterthought; it is built directly into the Pub/Sub specification, mirroring the robust security of client-server OPC UA. The framework provides message-level security for confidentiality and integrity, independent of the transport layer. Key features include:

  • Signing & Encryption: Messages can be digitally signed to prevent tampering and encrypted to prevent eavesdropping, using industry-standard algorithms.
  • Key Management: A dedicated Security Key Service (SKS) distributes symmetric keys to authorized Publishers and Subscribers, ensuring only trusted nodes participate in the data exchange. This allows secure multicast even on shared Ethernet networks.
04

DataSetMetaData & Self-Description

To enable true plug-and-produce interoperability, Pub/Sub uses DataSetMetaData. Before sending raw data, a Publisher transmits a structured description of the data payload, including field names, data types, and engineering units. Subscribers use this metadata to dynamically interpret incoming messages without needing a pre-configured, static schema. This self-describing mechanism is critical for semantic interoperability, allowing a robot from Vendor A to understand a sensor from Vendor B without manual mapping, drastically reducing commissioning time.

05

High-Performance Data Encoding

For applications requiring minimal latency and maximum throughput, such as motion control, Pub/Sub specifies highly efficient binary encodings. UADP uses a compact binary format that avoids the overhead of XML or JSON. It supports raw encoding, where the payload is a direct binary dump of the system's memory structure, eliminating serialization overhead entirely. Combined with TSN, this enables deterministic delivery guarantees in the sub-millisecond range, making it suitable for hard real-time control loops that were previously the domain of proprietary fieldbuses.

06

Broker-less Multicast with TSN Integration

A defining feature for industrial automation is the ability to operate without a broker. Using UADP over UDP multicast, Publishers can send a single message that is simultaneously received by all interested Subscribers on the network segment. When combined with Time-Sensitive Networking (TSN), this becomes a deterministic, lossless communication channel. TSN guarantees bounded low latency and zero congestion loss, enabling the replacement of legacy hard-wired signal lines with a converged Ethernet network that carries both real-time control data and best-effort traffic.

OPC UA PUB/SUB

Frequently Asked Questions

Clear, technical answers to the most common questions about the publish-subscribe extension for OPC Unified Architecture, designed for engineers and architects evaluating scalable industrial data distribution.

OPC UA Pub/Sub is an extension of the OPC Unified Architecture that enables scalable, broker-less, one-to-many or many-to-many data distribution using a publish-subscribe communication pattern. Unlike the client-server model, where each subscriber must establish a dedicated connection to the publisher, Pub/Sub decouples senders and receivers. A Publisher sends messages to a logical channel (a Topic), and any number of Subscribers listening to that topic receive the data without the publisher needing to know who they are. This is achieved by mapping OPC UA data to transport protocols like UDP multicast for local factory networks, AMQP or MQTT for cloud integration, or Time-Sensitive Networking (TSN) for deterministic, real-time delivery. The core mechanism involves configuring DataSet objects that define the payload structure, DataSetWriter objects that control the publishing behavior, and DataSetReader objects on the subscriber side that reconstruct the information model. This architecture dramatically reduces network overhead and configuration complexity in systems with hundreds of sensors broadcasting to multiple consumers.

COMMUNICATION PARADIGM COMPARISON

OPC UA Client-Server vs. Pub/Sub

Architectural and operational differences between the traditional client-server model and the publish-subscribe extension in OPC UA, critical for selecting the appropriate data distribution pattern for industrial automation workloads.

FeatureClient-ServerPub/SubPub/Sub with Broker

Communication Pattern

Point-to-point, request-reply

One-to-many, unidirectional

Many-to-many via intermediary

Connection Model

Stateful session with TCP

Connectionless (UDP) or connection-oriented

Broker-managed sessions

Discovery Mechanism

FindServers/GetEndpoints services

Multicast DNS-SD or LDS-ME

Broker topic enumeration

Data Encoding

UA Binary or XML

UA Binary, JSON, or UADP

Broker-native format with UA payload

Transport Protocols

UA TCP (opc.tcp://)

UDP multicast, Ethernet TSN, AMQP, MQTT

AMQP, MQTT 3.1.1/5.0

Subscriber Filtering

MonitoredItem filters on server

DataSetClass and ContentFilter at subscriber

Broker topic routing with server-side filtering

Security Model

Session encryption (UA Secure Conversation)

Message-level security (JSON Web Token, X.509)

Broker TLS + UA payload encryption

Quality of Service

Guaranteed delivery via TCP

Best-effort (UDP) or broker QoS levels

Broker-defined QoS (0, 1, 2)

Typical Use Case

Supervisory control, HMI data access

High-speed sensor multicast, TSN networks

Cloud integration, cross-site telemetry

Scalability Ceiling

Limited by server connection pool

Unlimited subscribers per publisher

Broker clustering limits

Prasad Kumkar

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.