Inferensys

Glossary

Address Space

A collection of Nodes and References that an OPC UA Server makes visible to Clients, representing a standardized, object-oriented view of underlying real-time data and system capabilities.
Data scientist building training data pipeline on laptop, data preprocessing visible, technical workspace.
OPC UA CORE CONCEPT

What is Address Space?

The Address Space is the standardized, object-oriented data model that an OPC UA Server exposes to Clients, representing all available information as a connected graph of Nodes and References.

An Address Space is a collection of Nodes and References that an OPC UA Server makes visible to Clients, providing a standardized, object-oriented view of underlying real-time data, historical data, alarms, and system capabilities. It is the single entry point through which a Client browses, discovers, and interacts with every piece of information the Server manages.

Nodes represent objects, variables, or methods, while References define the typed relationships between them, forming a rich semantic graph. This structure allows an Address Space to mirror complex physical systems—such as a production line—enabling a Client to navigate from a factory node down to a specific sensor value using a hierarchical, self-describing Information Model.

STRUCTURAL PRINCIPLES

Key Characteristics of the Address Space

The OPC UA Address Space is not a flat list of tags but a richly interconnected, object-oriented graph. Its power lies in how it models real-world assets, their relationships, and their semantics in a way that both machines and humans can navigate.

01

Object-Oriented Structure

The Address Space models data as a network of Nodes rather than a simple flat list. Each Node represents a real-world entity like a device, a sensor, or a process. Nodes are defined by their Attributes (properties like Value, DataType, and Timestamp) and connected by References (typed relationships). This allows a TemperatureSensor Node to be a child of a Boiler Node, inheriting context and making the system's structure self-describing.

02

Semantic Type System

Every Node is an instance of a specific NodeClass, which defines its fundamental behavior:

  • Object Nodes: Represent physical or logical system components (e.g., 'Pump_01').
  • Variable Nodes: Hold actual process values (e.g., 'Current_Speed').
  • Method Nodes: Represent callable server functions that a Client can invoke (e.g., 'EmergencyStop').
  • View Nodes: Provide filtered, task-specific subsets of the full Address Space, simplifying client navigation for specific user roles.
03

Rich Interconnectivity via References

Relationships are first-class citizens. References are not just parent-child hierarchies; they are typed pointers that create a semantic web. Common reference types include:

  • Organizes: Defines a standard hierarchical structure for browsing.
  • HasComponent: Indicates a part-of relationship (e.g., a Motor is a component of a Conveyor).
  • HasTypeDefinition: Links an instance Node to its template (Type Node), enabling instant discovery of its full semantic contract. This graph of typed connections allows a Client to traverse from a production line down to a specific sensor's alarm state.
04

Standardized Information Models

The base Address Space is extended by Information Models, which are formal, pre-defined schemas. The OPC UA specification provides core models, but industry-specific Companion Specifications (e.g., for Robotics or Machinery) add domain semantics. This means a 'Robot' Node from Vendor A and Vendor B will expose the same standardized structure and identifiers, enabling true plug-and-produce interoperability without custom mapping for each device.

05

Bidirectional Discovery

The Address Space is designed for dynamic exploration. A Client can browse the hierarchy to discover unknown Nodes and their relationships. Crucially, it can also query in reverse: a Client can ask, 'Which other Nodes reference this specific Node?' This bidirectional traversal is critical for impact analysis, allowing a system to instantly identify all processes that depend on a failing sensor by following incoming HasComponent or HasEffect references.

06

Stateful Eventing and Alarming

The Address Space is not just for static data. It integrates a sophisticated Alarms and Conditions (A&C) model. Alarm states (e.g., 'HighTemperature') are represented as Nodes with specific state machines. A Client can browse the Address Space to find all currently active alarms, their severity, and the specific Node they are associated with. This stateful model supports acknowledgment, confirmation, and shelving workflows directly within the information structure.

OPC UA ADDRESS SPACE

Frequently Asked Questions

Quick answers to common questions about the structure, mechanics, and practical use of the OPC UA Address Space for industrial interoperability.

An OPC UA Address Space is a collection of Nodes and References that an OPC UA Server makes visible to Clients, representing a standardized, object-oriented view of underlying real-time data and system capabilities. It works by modeling every piece of data, metadata, and system relationship as a Node with a unique NodeId in a graph structure. A Client browses this graph starting from the Root Node, following hierarchical Organizes references and non-hierarchical HasComponent or HasProperty references to discover available data. The Address Space is not a flat list of tags; it is a rich, typed information model where a temperature sensor is not just a value but a Variable Node with engineering units, a data type, and a reference to its parent device. This allows a Client to understand the meaning of data without prior configuration, enabling true plug-and-produce interoperability.

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.