An idea before a name
The notion that everyday objects could sense things and share what they sense is older than the term for it. In the late 1980s and early 1990s, researchers working on what they called ubiquitous computing imagined computers that would fade into the background of ordinary life, embedded in walls, tools and furniture rather than sitting on a desk. Mark Weiser of Xerox PARC is the name most often linked with that vision.
Imagining it was one thing. Making it work required cheap sensors, small processors and, above all, a network that objects could join.
Early connected objects
Some of the first examples were playful and practical at once. At Carnegie Mellon University in the early 1980s, students reportedly connected a soft drink vending machine to the campus network so they could check remotely whether it held cold bottles before making the trip. It is often cited as one of the first internet-connected appliances. Around the same time the network itself, descended from ARPANET, was spreading through universities.
In the early 1990s, researchers at the University of Cambridge pointed a camera at a shared coffee pot and put the images on the local network, so colleagues could check whether coffee was available. When the World Wide Web arrived, the pictures were made available to a wider audience. These projects had little commercial purpose, but they showed that a physical thing could report its state to anyone on a network.
Giving it a name
The phrase “Internet of Things” is generally credited to Kevin Ashton, who used it around 1999 while working on tracking goods with radio-frequency identification (RFID) tags. The thinking was that if objects could identify themselves and be read automatically, businesses would not depend on people typing in data, and could know what existed and where. The early emphasis was therefore on supply chains and inventory rather than homes.
Through the 2000s the idea spread into industry. Sensors monitored machinery, pipelines, shipping containers and farms. Wireless sensor networks, a research area of their own, explored how many tiny, battery-powered nodes could relay readings with little energy.
What made it affordable
Several shifts, arriving gradually, moved connected objects from specialist use to shops.
- Cheaper chips and sensors. Components developed for mobile phones, such as motion sensors, radios and small processors, became inexpensive and small. A device could gain a wireless chip for a modest extra cost.
- Wi-Fi in homes. As broadband and home Wi-Fi became normal, households already had a network for gadgets to join, so no special infrastructure was needed.
- Smartphones as controllers. The phone provided a screen, a keyboard and a connection, so a device did not need its own interface. An app could be the remote control for almost anything.
- Cloud services. Cloud computing allowed makers to store data, run apps and push updates without each customer running a server. It also made subscriptions and remote access possible, and shifted a lot of control to the company.
- Low-power wireless standards. Technologies such as Bluetooth Low Energy, Zigbee and later Thread let small battery devices run for long periods.
Reaching the home
Consumer products appeared in waves. Early connected bulbs, plugs and cameras came in the early 2010s. Around 2011 a learning thermostat popularised the idea of a connected device that homeowners actually noticed in daily life. Fitness bands and watches brought a related set of wearable devices to wrists. And in the middle of that decade, voice-controlled speakers made the smart home understandable to ordinary buyers: you could simply tell it what to do.
Ahead of those products, however, was a mess of incompatible systems. Each maker had its own app, protocol and cloud, a problem that later standardisation efforts, including the Matter standard, tried to ease. Meanwhile the number of connected products grew rapidly across homes, cars, cities and workplaces, though the figures quoted in forecasts varied widely and should be treated cautiously.
What the history shows
Looking back, a few patterns are visible.
The use came after the technology. The first connected objects solved small, local annoyances, such as checking a machine or a pot. Whether a use is truly valuable has often been settled only after products reached the public, and some connected gadgets turned out to add little over their plain equivalents.
Security and privacy lagged behind. As devices multiplied, many shipped with weak default passwords or poor update support, and reports of compromised cameras and routers became common. Connected products also created new streams of personal data, a topic now discussed under device telemetry. Makers and regulators in several places have since moved towards stronger baseline requirements, although practice still varies by product.
Longevity is a real issue. A conventional lamp can last decades. A connected one depends on an app, a service and a company that may change or disappear. That mismatch is now one of the most discussed practical downsides.
Where it stands today
The Internet of Things is no longer a single idea but a general description of a very large range of products, from industrial sensors to doorbells. For most households its presence is mundane: a thermostat that follows a schedule, a speaker that answers questions, a plug that switches a lamp. What began as a student project to avoid a wasted walk to a vending machine has become ordinary infrastructure, with all the convenience and the housekeeping that implies.