If you have a business or organisation which is reliant on a complex IT system for its output, or a product which runs on computers and associated infrastructure, how can you find out what would happen if one of those items of hardware or software became damaged or stopped working?
The answer is by using a digital twin.
We sat down with MAIAR Director Tony Reeves to quiz him about what digital twins are, how they help organisations make critical decisions, and what businesses should consider before investing.
Tony, in your own words, what is a “digital twin”?
So, in essence, a digital twin is a representation of something. That could be a system, enterprise, facility, network, or even a product like a car.
A digital twin should be designed to a level of fidelity that allows your organisation to test how the original would “react” in certain situations and then make decisions based on the outcomes – without impacting the “thing” that exists in the real world.
So, if you worked in the shipping industry, you might have a digital twin of a port because you need to understand how to maximise the throughput of vessels through the various berths: a digital twin would help you exercise that.
What does a digital twin help organisations do, and what problems do they solve?
So the first thing is that by building something that represents what actually exists, rather than what your documents might say exists, you’ve got something accurate to work from.
Digital twins give you real insights into whether what you think you have and what you do have are the same thing. You can capture the reality, and then compare that to what should exist so you can exert governance and control over whatever it is you’re representing.
Secondly, you can test hypotheses and “what-ifs” on a model if you’ve got a digital twin. And why do you have a model? So that you don’t have to touch the real “thing” in the real world.
In some cases, you just can’t test out those scenarios in the real world. You don’t want to be testing what would happen if a valve fails on a high-pressure hydrocarbon line coming from the North Sea because if there’s an issue with the test, you’re going to get a rather big bang! But, a model would allow you to safely test that potentially volatile and dangerous scenario.
The next stage of how digital twins help organisations is that they can be used for wargaming, resilience testing, and “what-iffing”! So you can use a twin to answer questions like “What if this particular component goes offline?”, “What if a cyber attack targets this system?”, “What if this aircraft were to enter the wrong flightpath?” or, returning to the port analogy, “What if a ship breaks down and blocks the berth?”. You can use the digital twin to model the impacts of those “what ifs” and understand the dependencies. And once you know which parts of your network depend on others, you can put plans and contingencies in place to mitigate damage in case those “what ifs” ever come to fruition.
The art is in deciding what fidelity level you need. Some people think digital twins are just walk-through, three-dimensional models of airport terminals showing the flow of people through the infrastructure. That could be true, but if your organisation is only looking at making decisions on an operational level about how many aircraft pass through each day, then you probably don’t need to worry about modelling individual people and accurately replicating the aesthetics of the inside of the terminal.
So, to make the most of a digital twin, you’ve really got to understand what answers your organisation is looking for.
Are there any kinds of organisations where using digital twins would be particularly useful/ beneficial?
There’s applicability in a whole host of areas!
Any organisation that has a complex IT enterprise should already be using this type of approach, and if they’re not, they should look at adopting it ASAP.
The NHS; national infrastructure organisations such as rail, and waterways; government departments; oil and gas businesses; complex organisations with multiple networks and systems working simultaneously; businesses which are dispersed geographically; and companies undergoing large-scale change projects would all benefit from using digital twins.
What are organisations “missing out on” by not using digital twins?
In simple terms, firstly, if organisations are conducting tests in the real world, they run the risk of breaking something that takes longer to recover than they had anticipated.
Secondly, they could inadvertently introduce a vulnerability that they weren’t expecting to. For instance, taking a security component offline to test it could result in a cyber attack.
Why do you think more organisations don’t use digital twins currently?
Digital twins have numerous benefits, but they can also be very costly. There are usually a lot of upfront costs, but the benefits can be delivered over a long period – so there can be quite a long return on investment.
They can be labour-intensive and take a long time to set up (whether doing the work yourself or outsourcing it to a company like ours that uses Sparx EA to create digital twins).
And I think there’s a lack of market awareness regarding what these digital twins can do for organisations. They’re a decision tool, and I think that’s the bit that the organisations don’t yet understand: that there is an amazing tool here that gives them the ability to make decisions without impacting on their real-world item.
Are there any risks/ downsides to using digital twins?
There are a couple, but not many.
If you don’t have a digital twin at the right level of fidelity for what your organisation needs, then the answers it gives might not necessarily be true. If you base a contingency plan around incorrect answers from a digital twin and then have to enact that plan in the real world, it might not work.
The ongoing maintenance you must do to ensure your digital twin remains accurate can be quite labour-intensive. Once the twin is up and running, the resources required are much less than the initial set-up – but, just like the real-world item this represents, it does require maintenance.
And the final downside I’d say organisations need to be aware of is that you need experts to be able to teach your people to use the twin correctly for it to give you the answers you need. You need to know what to input, how to prompt it, and what to ask it (which again is linked to understanding what level of fidelity you need).
If you get that wrong at the beginning, and you end up spending two years mapping everything in your organisation down to individual laptops on desks, but the sort of resilience questions you’re going to ask are, “What happens to our customer delivery if the power goes off?”, then you’ve spent all of your money working the wrong part of the problem.
What should organisations consider before using digital twins?
There are a few critical questions that businesses need to ask themselves, and information they need to gather, before they embark on the process of using a digital twin:
- “Is this going to be worth the money?”. Fundamentally, it comes down to a cost-benefit analysis.
- “Do we have enough money?”. You need enough of a budget to carry you through a long-term project which could span several years.
- “What fidelity do we need?”. Organisations need to know what questions they need answers to.
- “What documentation do we already have?”. You need to gather as much information and material as possible about your existing IT infrastructure: we’ve seen examples where customer projects have been hugely delayed because they cannot find, for example, architectural designs of their IT systems.
- “Are we prepared to uncover some uncomfortable truths?”. We often find that in the process of digital twins being created, organisations find issues with governance and policies that they weren’t aware of. You need to be prepared to deal with those as they crop up.
Can MAIAR help companies who are considering using digital twins?
Absolutely. Please get in touch if you’d like to find out more about our modelling services. We always take a consultative approach to projects like this and work with organisations to identify whether digital twins are a viable solution for them and, if so, what level of fidelity is required. If it’s not right for you, we’ll tell you.