by Peter Thorne
@Cambashi_Peter
LinkedIn
Two concepts – connected devices and digital twins – are central to Internet of Things (IoT) and digital transformation projects, yet they offer almost exactly opposite benefits. Here, Peter Thorne discusses how the opposing characteristics work together to provide a better foundation for connected applications.
Connected Devices vs Digital Twins
The concept of connected devices is not only at the heart of every IoT initiative, it also has the substantial benefit of being fairly straightforward to understand. If this device is connected, it must be able to send and/or receive information to/from other connected systems.
The concept of digital twin is also vital. This is the data model which enables the software in connected applications to provide useful functions. This is more abstract. What is this model? What can you do with it? What has it got to do with connected devices?
The first step to answering these questions is to appreciate that connected devices and digital twins offer almost diametrically opposed benefits. Connected devices enable real-world information to flow into the connected applications. Digital twins connect this world-as-it-is data to computer models which represent the world-as-in-theory-it-ought-to-be. Together these concepts make it possible to integrate the virtual world and the real world. Each is reaching out to the other, building the bridges which useful connected applications need.

Connected Devices
Connected devices enable visibility of data which is sensed from the real world. This is not a new concept – after all, retail point-of-sale systems have for decades used bar-code readers (sensors) to determine which real-world objects must be paid for. But this is somewhat unusual – many IT systems are blind to the real world, they depend on the data held in their databases, even if a ‘data entry’ activity in the past has loaded these databases with real world information. Perhaps an extreme example is the census in the UK – every 10 years, real world data is collected from the population and used to calibrate and refine forecasts and estimates used by local, regional and national government for the next 10 years.
Digital Twins
Digital twins are computer systems which can make sense of real world data coming from connected devices. Their designers enable this by finding ways to model the part of the real world that is within the scope of the digital twin.
In some cases, the reason a digital twin exists is as simple as to give easy, organised access and use of data sensed from the real world. The data representation of the contents of a shopping cart is a ‘digital twin’. The checkout software integrates pricing to calculate the total cost, including application of special offers. A digital twin fed by roadside traffic speed sensors enables a connected application to highlight the traffic jams on a map.
But the real power of digital twins comes from the next level, in which they help answer what-if questions. To answer what-if questions in the real world is expensive and cumbersome – you have to set up the conditions of your what-if question, and measure what happens. Not exactly convenient if the question is the relative costs and benefits of twenty 3-ton capacity trucks versus ten 30-ton trucks for your distribution business.
Of course people have found a myriad of ways to model the real world in order to answer what-if questions without resorting to try-it-and-see. From numeric models on spreadsheets to physical models of buildings used by architects, the diversity of the types of model, and the range of what-if questions they help answer, is enormous. Many computer systems contain these models – from distribution management to planning the layout of an office. These systems contain data about the real world, but in most cases the data they hold is not sensed directly, it is a separate representation which is estimated, entered and maintained to satisfy the purpose of the system.
Digital twins offer a new approach. The word ‘twin’ embodies the idea that the connected computer systems are no longer blind to the real world, they can integrate relevant data from connected devices.
How they work together in IoT
The architect of a successful IoT initiative will use these two concepts. Starting in the real world, connected devices create the opportunity to replace the estimated, modelled or manually entered data in a traditional computer application with real world data. Starting in the computer system, digital twins will reach out to connected devices to absorb and organise real data, and use it not only to report real status, but also support what-if questions.
And what-if questions are just the beginning. Most connected devices will be built to accept commands and updates from trusted IT systems, so the connected application be extended with control capabilities.
This view of connected devices and digital twins helps throw light on the reason for the growth of ‘simulation’ as a relevant technology for connected applications. To handle what-if questions, the connected application must contain a simulation model which covers the scope of the what-ifs. Naturally, the types of simulation model will be very different according to application area (a supply chain what-if is very different to a 3D-printer’s temperature gradient what-if). But all of these models, when connected to relevant real world data, deserve the ‘digital twin’ label.