A Win for Digital Twins

The Digital Twin Consortium is formed

I recently caught up with Dr Richard Soley, chairman and CEO of the Object Management Group, who has spent the past few months helping to set up the Digital Twin Consortium, announced on 18 May 2020. Founding members of the consortium include Ansys, Dell, Lendlease and Microsoft, with support at launch from a further 15 companies covering a range of areas.

The purpose of the group is to influence much-needed standardization in digital twin technology, and provide guidance, reference implementations and open-source software to enable the market to flourish.

Digital twins are a major concept in the Internet of things (IoT). They’re a virtual replica of a physical “thing”, useful for monitoring performance, simulating new set-ups (for example, machine learning for optimizing a process in a factory), carrying out analytics and testing new software builds including updates of machine learning models.

A digital twin can have several layers, including the identity of the thing; a data model, data from its real-world sibling and permissions for sharing data; metadata about the object, including compliance requirements; roles and policies for how the object behaves within a bigger system; a copy of the software running on it, the software build state and life cycle management; and APIs for interacting with it.

Much of the focus of early work in digital twins has been in manufacturing, thanks to projects like the Industrie 4.0 initiative led by German companies, but opportunities clearly could apply to any system. A digital twin can be set up for a single thing, or it can also be part of a composite digital twin: a virtual representation of a whole system of things at any scale, ranging from a car, a smart home, factory or building, to a power grid or telecom network for a whole country.

Having started out with no standardization, the market for digital twins is made up of a confusing mix of proprietary approaches, with a high degree of supplier lock-in, no common language or data standards and no interoperability. So today digital twins are used mainly by large engineering companies and some building information management systems, but they haven’t spread into the mainstream IoT market. This is what the Digital Twin Consortium is trying to achieve.

Dr Soley is realistic about the ambition and challenges that come with this aim. He’s clear that it’s early days and so not everything about the consortium is fully defined, including the boundaries on its activities. However, the timeline for the first year is well laid out and includes work to:

• Develop a library of use cases
• Define an open-source software implementation
• Define terminology and a taxonomy
• Identify different types of digital twin, such as those used in process industries and simulations
• Identify the product life cycle
• Find out which areas can be common between industries or specific to a vertical market.

Dr Soley has clear criteria to score early success with the consortium. There should be a shared language for digital twins, possibly with industry-specific layers to address compliance with regulatory requirements; clarity on use of the technology in at least four target vertical markets, including healthcare in the wake of the Covid-19 pandemic; some good open-source software implementations; as well as interoperability and portability of digital twins between systems.

The last of these criteria is a clear attempt to reduce supplier lock-in and customer risk. A core belief of the consortium is that the digital twin itself shouldn’t be a source of differentiation for a supplier. Instead it should focus on other factors such as quality, price, speed, performance and interworking with other systems. At first sight, portability seems a high bar to set, but it’s clearly needed. If a customer changes key systems in its operations, it will want to port twins over to the new system. Similarly, when there’s a merger or acquisition, it should be possible to integrate systems across the two companies.

Although it faces challenges, the consortium has launched at a good time and the initiative seems to have touched a nerve. At its Build 2020 conference on 19 May, Microsoft announced a large push into this area, expanding its Azure Digital Twin service to enable users to model country-scale systems. Microsoft is a founding member of the consortium, which has grown to 56 members after only 10 days in existence.

As the market for IoT systems grows and matures, several aspects of the systems are becoming open and horizontal, rather than just being part of an individual supplier’s vertical stack. This is an essential step toward achieving real scale in the industrial IoT market. The Digital Twin Consortium is an important and timely initiative that’s accelerating this effort for another layer in IoT.