Inferensys

Glossary

Information Model

A formal, object-oriented schema that defines the structure, relationships, and semantics of Nodes in an OPC UA Address Space, enabling machines to understand the meaning of data.
Data scientist building training data pipeline on laptop, data preprocessing visible, technical workspace.
OPC UA ADDRESS SPACE SCHEMA

What is an Information Model?

A formal, object-oriented schema that defines the structure, relationships, and semantics of Nodes in an OPC UA Address Space, enabling machines to understand the meaning of data.

An Information Model is a formal, object-oriented schema that defines the structure, relationships, and semantics of Nodes within an OPC UA Address Space, transforming raw data into machine-understandable information. It specifies the allowed types, instances, and constraints, enabling a server to expose a standardized, self-describing view of underlying system capabilities to any connected client.

Built using the OPC UA meta-model, an Information Model leverages object-oriented concepts like inheritance and composition to create domain-specific type hierarchies. Industry groups codify these as Companion Specifications, ensuring semantic interoperability where a robotic arm from one vendor and a vision system from another can exchange meaningful, context-rich data without custom integration code.

OPC UA SEMANTIC FOUNDATION

Key Characteristics of an Information Model

An OPC UA Information Model is the formal, object-oriented schema that transforms raw data into meaningful, machine-interpretable information. It defines the types, relationships, and constraints governing Nodes within the Address Space.

01

Object-Oriented Type System

Information Models use a strict object-oriented paradigm with Object Types and Variable Types that support inheritance hierarchies. A BaseObjectType defines common attributes, while specialized subtypes extend it with domain-specific properties.

  • Subtyping: A 'Motor' type inherits from 'Equipment', adding motor-specific variables like RPM and Torque.
  • Instantiation: Each physical motor on the factory floor becomes an Object instance of the Motor type in the Address Space.
  • Encapsulation: Types bundle data (Variables), behavior (Methods), and state (Alarms) into single, reusable definitions.
02

Semantic References

Nodes are connected via References, which are themselves typed and carry explicit semantic meaning. Unlike simple pointers, a Reference defines why two Nodes are related.

  • HasComponent: Indicates a part-of relationship (e.g., a Motor HasComponent a Drive).
  • HasProperty: Links an Object to a characteristic that cannot have its own children (e.g., a Motor HasProperty SerialNumber).
  • Organizes: Defines a hierarchical grouping for browsing, not a structural dependency.
  • HasCause: Links an Alarm to the Variable that triggered it, enabling root cause analysis.
03

Standardized Base Model

Every OPC UA Server starts with the OPC UA Base Information Model, a mandatory, pre-defined set of types in namespace 0 (http://opcfoundation.org/UA/). This provides the foundational vocabulary for all other models.

  • Defines core types: BaseObjectType, BaseDataVariableType, PropertyType.
  • Establishes the Address Space structure itself, including the Root, Objects, Types, and Views folders.
  • Provides the base Event and Alarm model for stateful condition monitoring.
  • All vendor or industry-specific models extend from this common root, ensuring baseline interoperability.
04

Companion Specifications

Industry working groups publish Companion Specifications—standardized, domain-specific Information Models that enable plug-and-produce interoperability across vendors.

  • OPC UA for Robotics: Defines types for RobotController, MotionDeviceSystem, and TaskProgram so any compliant robot exposes the same interface.
  • OPC UA for Machinery: Standardizes machine tool identification, job management, and condition monitoring.
  • Weihenstephan Standards: Models for packaging and food & beverage equipment.
  • These specifications are stored and versioned in the OPC UA Cloud Library for global access and consistency.
05

Instance Declaration vs. Modelling Rule

When defining a type, you specify Instance Declarations—placeholders for child Nodes that will exist on every instance. Modelling Rules dictate whether these declarations are mandatory or optional.

  • Mandatory: Every instance of the type must have this child Node (e.g., every Motor must have a CurrentSpeed Variable).
  • Optional: An instance may include this child Node (e.g., a Motor may have a CoolantTemperature Variable if equipped with a sensor).
  • This mechanism allows Clients to discover the guaranteed structure of any Object by examining its type definition, enabling robust, generic client applications.
06

Namespace Segregation

Information Models are partitioned into Namespaces identified by unique URIs, preventing naming collisions between different standards and vendors.

  • Namespace 0: Reserved for the OPC UA Base Model.
  • Namespace 1: Typically used for a server's local, vendor-specific extensions.
  • Custom Namespaces: Each Companion Specification or vendor model gets its own URI (e.g., http://yourcompany.com/UA/Machines).
  • A Node's NodeId combines a Namespace Index and an Identifier, ensuring global uniqueness across the entire Address Space.
Information Model

Frequently Asked Questions

Clear answers to the most common questions about OPC UA Information Models, their structure, and their role in enabling semantic interoperability in industrial automation.

An OPC UA Information Model is a formal, object-oriented schema that defines the structure, relationships, and semantics of Nodes within an OPC UA Server's Address Space. It works by transforming raw data points into meaningful, typed objects that machines can interpret without human intervention. Instead of exposing a flat list of tags like 'Temp_421', an Information Model exposes a 'Boiler' object with a 'Temperature' variable, an 'Alarm' condition, and a 'Start' method. This is achieved by defining ObjectTypes, VariableTypes, and ReferenceTypes that establish a strict hierarchy and inheritance. When a Client browses the Address Space, it discovers not just values but the meaning of those values—enabling plug-and-produce interoperability across vendors.

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.