Democratizing Digital Twins - A Management Platform for the Distribution and Sharing of Digital Twin Assets
Date
2024-05-02
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
ORCID
Type
Thesis
Degree Level
Masters
Abstract
Digital Twins (DT) - virtual models designed to accurately reflect physical objects, systems, or processes
- has received widespread adoption in its use to manage, monitor and respond to state changes of physical
objects or environments such as factories, health monitoring devices, smart homes, and even cities;
however, the primary focus of most research efforts have been on the challenges associated with sharing
of data generated by the DT assets, with less emphasis on the distribution and sharing of the DT models
themselves. This gap overlooks the access to models and the protection of intellectual property rights of the
creators.
This study introduces a management platform designed for the distribution and sharing of DT models,
facilitating control, personalization, and aggregation of DT Instances for specific objectives and generalized
control. This is achieved without compromising the embedded knowledge or user autonomy over data
sharing.
A key objective of this study is the exploration of the concept of Composite (Aggregate) Digital Twins –
the logical combination of several (atomic) Digital Twin Instances to facilitate the management and control
of a large number of heterogenous atomic DTs, without direct access to the instances themselves. By
representing the integration of several different Digital Twin Instances present in a Composite Digital Twin
as a JSON object, the study proposes to simplify the complex interaction of several interconnected, but
disparate, systems into a single, unified view.
This study presents the implementation of a Rust-based proxy/middleware web server, utilizing
webassembly-based containerized microservices, that allows for the deployment and hosting of DT models
packaged as Open Container Initiative (OCI) images - an open industry standards around container formats
and runtimes - and stored in repositories like Docker Hub in a vendor-agnostic environment. The services
exposed by these models are accessible via RESTful API endpoints and executable on container runtime
environments such as Docker Engine. The platform features a web portal for deploying and managing DT
models, enabling controlled sharing between model creators (Owners) and users (Subscribers) through
configurable access and data sharing policies.
Our evaluation focuses on the performance (latency and throughput) of accessing DT model services
directly versus through our middleware, highlighting the trade-offs in access costs for enhanced security,
data validation, and access control. The results affirm that the benefits of the middleware framework justify
the increased access costs – in terms of latency and throughput, for providing a secure and controlled
environment for sharing DT models and data.
Description
Keywords
Digital Twins, Containerization, Digital Asset Distribution and Sharing
Citation
Degree
Master of Science (M.Sc.)
Department
Computer Science
Program
Computer Science