What is the Internet of Things?

What does IoT cover?

A typical IoT application has some or all of the following characteristics:

  • The devices “in the field” (i.e. at the measuring or controlling points) have some or all of the following characteristics:
    • They are single-application – they can only do one job. E.g. a device that tells you if a door is open or closed. Note that normal computers or smartphones are the opposite as they can do many different jobs and are therefore not “single-application”.
    • They are cheap to produce. The electronic boards may be a few dollars.
    • They are numerous – thousands or more may be required to collect all of the data needed for the application. For example, to measure the air-quality over a wide area in a city and to determine blackspots, you need to be sampling the air at thousands of locations.
  • The devices don’t report directly to humans, but instead talk to other devices.
  • The devices don’t normally communicate directly with each other, and if they do, it’s typical that they only talk to close neighbours. Instead they normally report to a central bank of servers in either a public or private cloud. The method of this “backhaul” communication can vary but is often wireless – for example, it could be low-powered radio (LPR) or the mobile phone network,. The choice of comms technology will be determined by lots of factors like cost, availability, bandwidth requirements, impact on battery life.
  • The method of communications will use internet-style concepts, such as IP, HTTP, web servers etc. The same underlying technology that powers websites and email.
  • The data will be collected by software running on servers, and is likely to be held in perpetuity in some kind of database.
  • Applications will run against the data collected. As the data is likely to be vast, good applications will summarise and present the data in an easy-to-digest format. In my air quality example, the data from thousands of devices could be shown on a colour-coded map, and it could be possible to scroll forwards or backwards in time. Thus millions of pieces of data can be easily visualised and absorbed.

The above definition is quite flexible, it’s not definitive, and if an application displays some but not all of these characteristics it could also be argued that it’s still an IoT application.

Here’s a list of typical applications which are in existence now, or are slated for the near future:

  • All kinds of Smart City applications – smart lighting, air quality, traffic and parking management, noise, etc
  • Smart metering. Automatic meter reading and download of flexible tariffs. E.g. the UK Smart Metering Rollout.
  • Utilities – electricity, gas, water – network and asset monitoring and control
  • Fleet tracking – keeping track of where your vehicles are, what state they’re in, how they’re being driven etc
  • Home automation – security, appliance control


The IoT concept has a large overlap with some much older terms – for example “embedded systems” or “machine to machine (M2M) applications”. Both of these terms pre-date the advent of internet technology (web servers, HTTP etc), but they have the some of the same features as IoT, for example:

  • The start and end points in the system are usually both machines (rather than, for example, making a mobile phone call where the end points are both human)
  • The devices are usually small, numerous, cheap and for a single application.

The crucial difference is that IoT applications leverage internet technology building blocks, such as IP, HTTP, web servers etc, whereas an embedded system doesn’t necessarily have to use internet.

The term “Internet of Things” was first coined in 1999, whereas embedded devices have been around since the 1960s.


The IoT has a few key challenges that need to be addressed for most applications:

  • Like any computer system, an IoT application can have the potential to be hacked. This can be mitigated in many of the usual ways, but there are additional challenges for IoT devices, such as the devices being limited in their capability, and the challenges of distributing patches to many devices. There have been a few high-profile IoT hacks already.
  • As more of our activity is monitored and recorded (e.g. if your home security system records who enters and leaves your property, and when), there is the possibility that our privacy will be compromised and possibly used against us in an undesirable way.
  • IoT applications often use many devices in the field. The devices will all run software which will inevitably require patching over time. The devices may be battery-powered, and the batteries will have to be renewed. Devices may fail of their own accord, or be damaged by other activity in the vicinity. All of these aspects, and more, result in the need to manage the estate of devices. Inevitably, at any given time, not all devices will be providing good quality data; the IoT system must be able to tolerate this data degradation.
  • There are quite a few different, competing standards for communications at the moment. Interoperability and future-proofing devices is an issue.
  • The “could v should” argument – just because you can make a device IoT-enabled, doesn’t necessarily mean you should. There are a multitude of devices where you could legitimately question the benefit of having them online.


The IoT, whilst only having existed for less than 20 years, is here to stay and will affect many aspects of our lives in the western world. Like all new technologies, its benefits are accompanied by some significant challenges; however, used in the right way, the IoT can help us become more efficient at managing our lives and our limited resources.