Pub-Sub Model

The defining characteristic of the Pub-Sub model is the decoupling of Publishers and Subscribers. A Publisher sends data to a Message-Oriented Middleware (broker) without knowledge of the number, identity, or location of Subscribers. Subscribers express interest in specific DataSets without establishing a direct session with the Publisher. This eliminates the tight binding of the Client-Server model, allowing systems to be added, removed, or updated independently without reconfiguring the entire network.

Unlike the connection-oriented Client-Server model, Pub-Sub is fundamentally message-centric. A Publisher assembles a DataSetMessage containing field-level data values and pushes it to the network as a self-contained payload. Key properties include:

• One-to-Many Distribution: A single message is efficiently fanned out to multiple Subscribers by the broker.
• NetworkMessage: The standardized OPC UA container that wraps one or more DataSetMessages for transport.
• Payload Agnosticism: The broker does not interpret the payload; it only routes based on topic strings, preserving end-to-end data integrity.

The OPC UA Pub-Sub model is transport-agnostic, decoupling the data model from the underlying network protocol. This allows architects to select the optimal transport for a given use case:

• OPC UA PubSub over MQTT: Uses a lightweight MQTT broker for cloud-to-edge integration, ideal for bandwidth-constrained or high-latency networks.
• OPC UA PubSub over TSN: Combines Pub-Sub with Time-Sensitive Networking to guarantee deterministic, bounded low-latency delivery for real-time control loops.
• OPC UA PubSub over UDP: Supports multicast UDP for high-speed, local-area data distribution without a broker.

Data is organized into DataSets—pre-defined collections of related field values configured in the Publisher. Each DataSet is mapped to a Topic on the broker, which acts as the routing address. Subscribers listen on specific Topics to receive updates. This structure enables:

• Semantic Grouping: Related variables (e.g., all sensors on a motor) are packaged together.
• Selective Subscription: Subscribers receive only the data they need, minimizing processing overhead.
• Standardized Encoding: DataSets are serialized using UA Binary or JSON Encoding before being embedded in NetworkMessages.

Pub-Sub security is implemented at the message level rather than the transport level. Each NetworkMessage can be individually signed and encrypted using a Security Policy, ensuring end-to-end confidentiality and integrity regardless of the intermediary broker's trustworthiness. Key mechanisms include:

• Message Signing: Guarantees the message originated from an authorized Publisher and has not been tampered with.
• Message Encryption: Protects payload confidentiality as it traverses untrusted network segments.
• Key Rotation: Supports periodic renewal of cryptographic keys without disrupting the data flow.

OPC UA FX (Field eXchange) extends the Pub-Sub model for high-performance, controller-to-controller communication. It standardizes deterministic data exchange by combining:

• Scheduled Publishing: Messages are transmitted on a strict, time-triggered schedule.
• Time-Sensitive Networking (TSN): Network infrastructure guarantees bounded latency and zero congestion loss.
• Fail-Operational Redundancy: Supports redundant publishers and network paths for safety-critical applications. This makes Pub-Sub suitable for replacing hard-wired I/O and proprietary fieldbuses with open, interoperable Ethernet-based communication.

The foundational OPC UA communication pattern where a Client initiates a Session to request data, browse the Address Space, and call methods on a Server. Unlike Pub-Sub, this model requires a direct, stateful connection between both parties.

• Connection: Point-to-point, session-based
• Coupling: Tight; client must know server endpoint
• Use Case: Supervisory control, configuration, historical data access
• Contrast: Pub-Sub decouples senders from receivers via a broker

A transport protocol mapping that uses the lightweight MQTT publish-subscribe broker to distribute OPC UA DataSet messages. This is the most common cloud and IoT integration pathway.

• Broker: MQTT broker (e.g., Mosquitto, HiveMQ)
• Encoding: Typically JSON for web compatibility
• QoS Levels: Supports MQTT Quality of Service 0, 1, or 2
• Advantage: Firewall-friendly, works over standard TCP port 1883/8883

A transport profile combining PubSub with Time-Sensitive Networking to guarantee bounded low latency and jitter. Essential for deterministic, real-time industrial communication.

• Latency: Sub-millisecond, guaranteed delivery
• Topology: Layer 2 Ethernet with IEEE 802.1Qbv scheduling
• Use Case: Motion control, high-speed I/O, controller-to-controller
• Key Feature: Eliminates non-deterministic network behavior

A defined collection of field-level data values configured in a Publisher and delivered as a single payload in a PubSub Network Message. DataSets are the atomic unit of publication.

• Structure: Defined by a DataSetMetaData describing fields
• Encoding: Can be UA Binary, JSON, or vendor-specific
• Versioning: Supports major/minor version for schema evolution
• Example: A DataSet might bundle temperature, pressure, and flow rate from a single machine

An extension of the PubSub model that standardizes field-level, controller-to-controller communication with deterministic data exchange for high-speed automation.

• Full Name: OPC UA Field eXchange
• Scope: Replaces hardwired I/O and proprietary fieldbuses
• Key Concept: Publisher/Subscriber groups with pre-configured connections
• Benefit: Vendor-agnostic, software-defined field-level networking

The Client-Server equivalent of PubSub's data delivery mechanism. A Monitored Item specifies a Node attribute to watch, and a Subscription groups items and controls notification pacing.

• Trigger: Value change, deadband exceeded, or timer
• Delivery: Server pushes notifications to Client
• Contrast with PubSub: Requires an active Session; PubSub uses a broker with no session affinity
• Use Case: HMI updates, alarm monitoring, trending