In the last step, we looked at devices as a duality of physical and digital. It’s not always a perfect way to view the devices but it is a useful abstraction.
Abstractions are essential for reasoning about complex systems. They prevent us from being bogged down by detail and the best ones let us reason about the abstraction with confidence that the reasoning applies to the concrete system underneath.
For example, I could discuss creating a new smart device with a colleague. Let’s say it was a smart lamp. The digital twin abstraction is useful because when we sit down to design the product, we can start with a physical lamp and work on its digital twin, and we will both know what that means. Otherwise I would have to describe an electronic platform associated with the physical device that is a mixture of implementations made use of by the hardware interface and software interfaces that are resolved using the physical components – it is possibly more accurate than ‘digital twin’ but it doesn’t really help. In fact, such a heavy description just gets in the way.
There are a few other common abstractions that are used when discussing IoT devices that are worth thinking about before we go any further.
Model and interface
This abstraction is used in other areas, such as software development. It divides the device into two parts: the model and the interface.
The model is the data and processing required to hold some kind of internal representation of how things are or should be. Imagine a thermostat. The model is made up of a number representing the current temperature, another representing the target temperature and some logic that dictates what action should be taken based on the difference between the two.
The interface is how that model is affected by the world and how it affects the world. Using our thermostat example, this is a combination of the physical box that a user can interact with using buttons and dials, and the connection to the heating and/or cooling system that can be used to change the current temperature in order to match the target.
© Coventry University. CC BY-NC 4.0