Defining a Digital Twin Platform
Digital Twin is considered one of the key trends for IoT. The idea of a digital twin is to create a digital replica of a physical object and use the twin as the main point of digital interaction. A digital twin could be a factory lathe, a windmill, a container ship, or automobile. Gartner Group has listed digital twins as one of their top 10 technology trends in 2018. A survey conducted by Gartner found close to half of organizations are using or plan to use digital twins in 2018.
What is less clear is the technology required to build and deploy digital twins. What technology is needed for organizations to make digital twins a reality for large-scale deployments. How do you manage thousands of digital twins or thousands of different types of digital twins? How do you integrate a digital twins with other systems?
For organizations that want to scale out a digital twin strategy, I believe they are going to require something like a digital twin platform. However, the question becomes what defines a digital twin platform? It seems there are 6 key features that would define any scalable digital twin platform.
- Manage the Digital Twin Lifecycle — How do you design, build, test, deploy and maintain a digital twin and its digital master? Digital twins are the instantiations of something called the digital thread or digital master. For instance, each windmill will be based on a common digital master that represent the engineering diagrams, bill of material, software versions and other artifacts used to create each windmill. A digital twin platform needs the tools to bring all this information together to create the digital master and then manage any changes to the digital master. Tools then need to be available to test, deploy and manage each digital twin based on a specific digital master. The tools also need to be able to handle hundreds of digital masters and thousands of digital twins.
- Single Source of Truth — A digital twin is suppose to be an exact replica of a physical asset and the data coming from the asset. However, the maintenance of a physical asset can often change its physical state. For instance, a specific windmill might have a replacement part or a different version of firmware installed; other windmills might not have the same updates. A digital twin platform is able to update and provide the exact state for each individual digital twin, creating a single source of truth.
- Open API — A well defined digital twin becomes the interface and integration point for an Industrial IoT solution. A digital twin platform provides an open api that allows any system to interact with the digital twin. For instance, machine learning and analytics services should be able to interact with a digital twin through an API. An organization should be able to integrate their digital twin into other enterprise systems, like ERP or SCM systems. A digital twin API should really enable the concept of Device as a Service. The Eclipse Ditto project provides a good illustration of an API for a digital twin.
- Visualization and Analysis — Organizations should be able to use their digital twin platform to create visualizations, dashboards, and in-depth analysis of this live data from the digital twin. The live data should be linked to the digital master to allow for drill-downs into design documents or other components of the digital master.
- Event and process management — It should be possible to setup events and business processed to be executed based on digital twin data. For instance, events to schedule maintenance calls based on live data and an accurate view of the current maintenance status of the physical asset.
- Customer and User Perspective — A digital twin platform needs to provide a customer and user perspective. What organization own or operate each digital twin? What users are allowed to access the data and information associated with the digital twin. How can you share information with other users that might be involved in the design of a new release for the asset. A digital twin platform needs to enable collaboration between the stakeholders of the digital twin.
This list is meant as a starting point for discussion. It does not seem like there has been much discussion on how to actually implement digital twins. I am interested in hearing what might be missing or what isn’t important.
Disclosure: I am consulting for a company that is interested in providing a solution for digital twins and IoT solutions.