แก้ไข

แชร์ผ่าน


What is an ontology?

This article describes the concept of industry ontologies and how they can be used within the context of Azure Digital Twins.

The vocabulary of an Azure Digital Twins solution is defined using models, which describe the types of entities that exist in your environment. An ontology is a set of models that comprehensively describe a given domain, like manufacturing, building structures, IoT systems, smart cities, energy grids, web content, and more.

When you author a model set from scratch that is complete and describes a domain, you're creating your own ontology. Alternatively, sometimes modeling standards for an industry already exist, and it can be more effective to lean on that existing ontology instead of creating the ontology from scratch yourself.

The articles in this section explain more about creating ontologies and using pre-existing industry ontologies for your Azure Digital Twins scenarios, including what existing ontologies are available today, and the different strategies for turning industry standards into ontologies for use in Azure Digital Twins.

Summary of ontology strategies for Azure Digital Twins

Here are the main strategies for creating DTDL ontologies to use in Azure Digital Twins. Choose the one that works best for you, depending on how closely the existing materials for your industry already match your solution.

Strategy Description Resources
Adopt You can jump-start your solution by adopting one of Microsoft's open-source DTDL ontologies that has been built on widely accepted industry standards. If the ontologies contain all the models you need, you can take these model sets as they are and use them out-of-the-box. Adopting industry standard ontologies
Extend If an existing DTDL ontology has most, but not all, of the models you need in your solution, you can extend the ontology with your own additions to create a customized ontology. Adopting industry standard ontologies

Extending ontologies
Convert If you already have existing models represented in another standard industry format, you can convert them to DTDL to use them with Azure Digital Twins. Converting ontologies
Author You can develop your own custom DTDL ontologies from scratch, using any applicable industry standards as inspiration. DTDL models

Using existing ontologies

Existing industry ontologies provide a great starting point for digital twin solutions. They encompass a set of domain-specific models and relationships between entities for designing, creating, and parsing a digital twin graph. Industry ontologies enable solution developers to begin a digital twin solution from a proven starting point, and focus on solving business problems.

Using these ontologies in your solutions can also set them up for more seamless integration between different partners and vendors, because ontologies can provide a common vocabulary across solutions.

Here are some other benefits to using industry-standard DTDL ontologies as schemas for your twin graphs:

  • Harmonization of software components, documentation, query libraries, and more
  • Reduced investment in conceptual modeling and system development
  • Easier data interoperability on a semantic level
  • Best practice reuse, rather than starting from scratch

Microsoft has created several open-source DTDL ontologies built on widely used industry standards. You can adopt these ontologies out-of-the-box in your solutions, or extend the ontologies with your own additions to customize your solutions. Because models in Azure Digital Twins are represented in Digital Twins Definition Language (DTDL), ontologies designed for Azure Digital Twins are written in DTDL.

If you have a set of models for your industry that's represented in a different standard industry format, such as RDF or OWL, you can use it as a starting point and convert the models to DTDL in order to use them in Azure Digital Twins.

Authoring your own ontologies

If there's no existing industry ontology that meets your needs, you can always develop your own custom DTDL ontologies from scratch. These can be inspired by applicable industry standards, or any other information that's relevant to your business.

For information about designing individual models, including all the fields they contain and how to author them in DTDL, see DTDL models.

Full model development path

No matter which strategy you choose for designing or integrating an ontology into Azure Digital Twins, you can follow the complete path below to guide you through creating and uploading your ontology as DTDL models.

  1. Start by reviewing and understanding DTDL modeling in Azure Digital Twins.
  2. Proceed with your chosen ontology strategy: adopt, convert, extend, or author your models based on the needs of your solution and industry.
  3. Validate your models to verify they're working DTDL documents.
  4. Upload your finished models to Azure Digital Twins, using the APIs or a sample like the Azure Digital Twins model uploader.

Once your models have been uploaded to the service, you can...

  • Visualize the models in your ontology using the model graph in Azure Digital Twins Explorer.
  • Manage them on an ongoing basis, including retrieving models in code, updating models, and deleting models, using the instructions in Manage DTDL models.
  • Use the models to create digital twins and a twin graph.

Next steps

Read more about the strategies of adopting, extending, and converting existing ontologies:

Or, learn about how models are used to create digital twins: Digital twins and the twin graph.