Reusable Assets
Reusable assets make it easier to work with digital twins, and reusability of assets is a fundamental feature of the platform[1].
Kinds of Reusable Assets
The DTaaS platform sorts all reusable library assets into five categories:
Data
The data sources and sinks available to a digital twin. Typical data sources are sensor measurements from Physical Twins and test data provided by manufacturers for calibrating models. Typical data sinks are visualisation software, external users and data storage services. Some outputs, such as events and commands, are akin to control outputs from a Digital Twin. These control outputs usually go to Physical Twins, but they can also go to another Digital Twin.
Models
Model assets describe different aspects of Physical Twins and their environment, at different levels of abstraction, so the same Physical Twin can have several models. For example, a flexible robot used in a car production plant may have structural models that help track the wear and tear of parts. The same robot can have behavioural models describing the safety guarantees provided by the robot manufacturer, and functional models describing what parts the robot can manufacture.
Tools
Tool assets are software used to create, evaluate and analyse models. These tools run on a computing platform, such as an operating system, a virtual machine like the Java virtual machine, or a docker container. Tools tend to be platform specific, which makes them less reusable than models. A tool can be packaged to run on a local or distributed virtual machine, so the most suitable execution environment can be chosen for a Digital Twin. Most models need tools to evaluate them against data inputs. In some cases, executable packages run as binaries in a computing environment. Each of these packages combines models and tools into a ready-to-use Digital Twin.
Functions
Functions pre-process and post-process data inputs, data outputs and control outputs. Data science libraries and functions can be used to create function assets for the platform. Some Digital Twin models need calibration before use; functions written by domain experts, together with the right data inputs, can calibrate them. Functions can also process the sensor and actuator data of both Physical Twins and Digital Twins.
Digital Twins
Ready-to-use digital twins created by one or more users. They can be reconfigured later for specific use cases.
File System Structure
Each user's assets are stored in five directories, one for each category above. In addition, common library assets are available to all users. A simplified example of the structure is:
Tip
The DTaaS is agnostic to the format of assets. The only requirement is that they are files that can be uploaded on the Library page. Directories can be compressed into single files and uploaded, then decompressed from a terminal or the xfce Desktop available on the Workbench page.
A recommended file system structure for storing assets is also available in DTaaS examples.
Upload Assets
Users can upload assets into their workspace from the Library page of the website.
Open a directory and click the upload button to upload files or directories into the workspace. The uploaded assets are then available in all workbench tools. The new dropdown menu creates new text-based files on the page. Other files must be uploaded with the upload button.
The user workbench provides the following services:
- Jupyter Notebook and Lab
- VS Code
- XFCE Desktop Environment available via VNC
- Terminal
Users can also bring DT assets into their workspaces from external sources with any of these services. Developers using git repositories can clone from and push to remote git servers. Users can also use common file transfer protocols such as FTP and SCP to bring DT assets into their workspaces.
References
[1]: Talasila, Prasad, et al. "Composable digital twins on Digital Twin as a Service platform." Simulation 101.3 (2025): 287-311.