Frequently Asked Questions

Abbreviations

Term	Full Form
DT	Digital Twin
DTaaS	Digital Twin as a Service
PT	Physical Twin

General Questions

What is the DTaaS platform?

The DTaaS platform is a software platform for creating and running digital twins. The features page gives an overview of what DTaaS can do.

What is the scope and current capabilities of the DTaaS platform?

1. DTaaS is a web-based interface for invoking the tools needed to work with one or more DTs.
2. DTaaS lets users run DTs in their private workspaces. These user workspaces are based on the Ubuntu 20.04 operating system.
3. DTaaS can help create reusable DT assets only if the DT asset authoring tools run in the Ubuntu 20.04 xfce desktop environment.
4. From the DTaaS platform's perspective, DTs are executables. Users are free to work with DTs in whatever way suits them. DTaaS suggests creating DTs from reusable assets and offers a suggested structure for DTs. The examples show the DTaaS workflow in more detail. This suggested workflow is not mandatory.
5. DTs can run as services with a REST API from within user workspaces, which makes service-level DT composition possible.

What can not be done inside the DTaaS platform?

1. DTaaS does not help install DTs obtained from external sources.
2. Windows-only tools cannot run inside DTaaS workspaces.

Is there any fundamental difference between commercial solutions like Ansys Twin Builder and the DTaaS platform?

Commercial DT platforms like Ansys Twin Builder integrate models, simulation and sensors tightly. This leaves fewer choices in DT design and implementation, and it brings vendor lock-in. DTaaS, on the other hand, lets users split a DT into reusable assets and combine these assets flexibly.

Is licensed software such as Matlab provided?

Proprietary and commercially licensed software is not available on the platform by default. However, each user has a private workspace based on a Linux xfce desktop environment. Users can install proprietary and commercially licensed software in their workspaces. A screencast shows Matlab Simulink running within the DTaaS platform. Licensed software installed by one user is not available to other users.

Digital Twin Assets

Can the DTaaS platform be used to create new DT assets?

The core purpose of DTaaS is to help users create DTs from assets already available in the library. However, users can use the services available in their workspace to install asset authoring tools there. These authoring tools can then be used to create and publish new assets. User workspaces are private and are not shared with other users. Thus, any licensed software tools installed in a workspace are only available to that user.

Digital Twin Models

Can the DTaaS platform create new DT models?

DTaaS is not a model creation tool, and it does not create digital twin models itself. It can still help users create them: model creation tools can be placed inside DTaaS, and any Linux desktop or terminal tool can run inside DTaaS. Thus, models can be created inside DTaaS and executed using tools that run on Linux. Windows-only tools cannot run in DTaaS.

How can the DTaaS platform help to design geometric model? Does it support 3D modelling and simulation?

DTaaS by itself does not produce any models. It provides a platform and an ecosystem of services that let digital twins run as services. Since each user has a Linux OS at their disposal, they can also run digital twins that have a graphical interface. In short, the DTaaS platform is neither a modelling nor a simulation tool. If such tools are required, they must be brought onto the platform. For example, if Matlab is needed, the user must bring the licensed Matlab software.

Can the DTaaS platform support only the information models (or behavioural models) or some other kind of models?

The DTaaS platform is agnostic to the kind of models used. DTaaS can run all kinds of models, including behavioural and data models. As long as the models and matching solvers can run on a Linux OS, they can be used in DTaaS. In some cases, models and solvers (tools) are bundled together to form monolithic DTs. The DTaaS platform can run such DTs as well. DTaaS does not provide dedicated solvers. If a solver can be installed in the workspace, the platform does not need to provide one.

Does it support XML-based representation and ontology representation?

Not at present. The team is looking for users who need this capability. Given concrete requirements and an example, the team can discuss how to realise it in DTaaS.

Communication Between Physical Twin and Digital Twin

How can the DTaaS platform control the physical entity? Which technologies it uses for controlling the physical world?

At an abstract level, data flows from the physical entity to the digital entity and back to the physical entity. The person designing the digital entity decides how this communication happens. DTaaS provides communication services that make this communication easier. The InfluxDB and RabbitMQ services hosted on DTaaS may be used for two-way communication between digital and physical entities. RabbitMQ has its MQTT plugin enabled, so it also works as an MQTT broker.

How is a physical entity such as shape, size, weight, structure, or chemical attributes measured using DTaaS? Is any specific technology used in this case?

The measurements are made at the physical twin and then sent to the digital twin. A digital twin platform like DTaaS can only facilitate the communication of these measurements from the physical twin. DTaaS provides InfluxDB and RabbitMQ (which also acts as an MQTT broker) for this purpose. These are among the most widely used services for digital twin communication. DTaaS also permits the use of other communication technologies and services hosted elsewhere on the Internet.

How can real-time data differ from static data and what is the procedure to identify dynamic data? Is there any UI or specific tool used here?

The DTaaS platform cannot tell whether data is static or dynamic. It can store names, units and any other text description of the quantities of interest (weight of batter, voltage output, etc.). It can also store the data sent by the physical twin. The user has to make the distinction between static and dynamic data. Only metadata can reveal more about the nature of the data. A tool may help in very specific cases, but it still needs metadata. If a person makes this distinction, the need for metadata goes down but does not go away entirely. Some manufacturer-supported DT platforms tightly integrate data and model. In such cases, the tool itself manages the metadata. DTaaS is a generic platform for running digital twins. If a tool runs on a Linux desktop or command line, DTaaS can support it. The tool (e.g., Matlab) can then manage the metadata requirements itself.

Digital Twin DevOps Automation

Can a DT execute forever?

The web UI imposes a 10-minute timeout on an execution. Users can also terminate an execution manually. The best option is for a DT to run a terminate script, which ends the execution and returns the logs.

Is the DT execution really scalable?

The capacity of the DT execution infrastructure depends on GitLab and on the compute power available to the GitLab runners. GitLab itself does not limit the number of runners.

If many GitLab runners are attached to a GitLab repository, which one is used?

This is indeterminate. The order and location of a DT's execution cannot be relied on.

Data Management

Can the DTaaS platform collect data directly from sensors?

Yes, via the platform services.

Does DTaaS support data collection from different sources like hardware, software and network? Is there any user interface or any tracking instruments used for data collection?

The DTaaS platform provides InfluxDB, PostgreSQL, RabbitMQ, MQTT, MongoDB and ThingsBoard services. Both the physical twin and the digital twin can use these services to communicate. IoT (time-series) data can be collected using the InfluxDB and MQTT broker services. InfluxDB has a user interface for analysing the collected data. Users can also upload their own data files into DTaaS.

Is the DTaaS platform able to transmit data to cloud in real time?

Yes, via the platform services.

Which transmission protocol does the DTaaS platform allow?

InfluxDB, RabbitMQ, MQTT and anything else that can be used from cloud service providers.

Does the DTaaS platform support multisource information and combined multi sensor input data? Can it provide analysis and decision-supporting inferences?

Information from multiple sources can be stored. The InfluxDB service hosted on DTaaS has the Influx / Flux query language for sensor fusion, analysis and inference.

Which kinds of visualisation technologies can the DTaaS platform support (e.g. graphical, geometry, image, VR/AR representation)?

Graphical, geometric and image visualisation. If the visualisation needs specific licensed software, the user must provide the licence for it. DTaaS does not support AR/VR.

Platform Native Services on the DTaaS Platform

Is the DTaaS platform able to detect the anomalies about-to-fail components and prescribe solutions?

This is the job of a digital twin. If a ready-to-use digital twin exists for this purpose, DTaaS lets others use it. Anomaly detection can be done with platform services such as Grafana, ThingsBoard and InfluxDB.

Comparison with other DT Platforms

All the DT platforms seem to provide different features. Is there a comparison chart?

Here is a qualitative comparison of DT integration platforms:

Legend: high performance (H), mid performance (M) and low performance (L)

DT Platforms	License	DT Development Process	Connectivity	Security	Processing power, performance and Scalability	Data Storage	Visualisation	Modelling and Simulation
Microsoft Azure DT	Commercial Cloud	H	H	H	M	H	H	H
AWS IOT Greengrass	Open source commercial	H	H	H	M	H	H	H
Eclipse Ditto	Open source	M	H	M	H	H	L	L
Asset Administration Shell	Open source	H	H	L	H	M	L	M
PTC Thingworx	Commercial	H	H	H	H	H	M	M
GE Predix	Commercial	M	H	H	M	L	M	L
The DTaaS Platform	Open source	H	H	L	L	M	M	M

Adapted by Tanusree Roy from Tables 4 and 5 of the following paper.

Ref: Naseri, F., Gil, S., Barbu, C., Cetkin, E., Yarimca, G., Jensen, A. C.,
... & Gomes, C. (2023). Digital twin of electric vehicle battery systems:
Comprehensive review of the use cases, requirements, and platforms.
Renewable and Sustainable Energy Reviews, 179, 113280.

All the comparisons between DT platforms seems so confusing. Why?

The confusion comes from the fact that different DT platforms (Azure DT, GE Predix) provide different kinds of DT capabilities. GE Predix can run all kinds of models natively, even next to (on) the PTs. The Azure DT service cannot do this natively. The user must integrate it with other Azure services or third-party services to get the capabilities that GE Predix provides natively in one interface. The right platform therefore depends on the specific requirements.

GDPR Concerns

Does the platform adhere to GDPR compliance standards? If so, how?

The DTaaS platform does not store any personal information about users. It stores only usernames to identify users, and these usernames do not contain enough information to deduce a user's true identity.

Which security measures are deployed? How is data encrypted (if exists)?

The multi-user server installation requires an HTTPS-terminating reverse proxy between the users and the DTaaS platform installation. Administrators can install HTTPS certificates into the application. The codebase can produce an HTTPS-enabled application, and users can also install their own certificates obtained from certificate authorities such as Let's Encrypt.

What security measures does the cloud provider offer?

The DTaaS platform can be installed on a corporate server hosted behind network firewalls, so that only permitted user groups have network access and physical access to the server.

How is user access controlled and authenticated?

The multi-user server installation of DTaaS has two levels of authorisation. The first level is server-side OAuth 2.0 authorisation through Traefik forward-auth over a secure HTTPS connection. The second level is the OAuth 2.0 PKCE authorisation flow in the web client. A GitLab instance provides both. DTaaS does not store the account and authorisation information of users.

Does the platform manage personal data? How is data classified and tagged based on the sensitivity? Who has access to the critical data?

The platform does not store personal data of users.

How are identities and roles managed within the platform?

The platform has two roles: administrator and user. The administrator manages the user roles.