A drinks machine on the network
One of the best-known early examples of a networked object is a soft drink machine. In the early 1980s, students and staff in the computer science department at Carnegie Mellon University connected a Coca-Cola vending machine to the department’s computers, which were linked to the early network that grew into the internet. Walking to a machine that might be empty or full of warm bottles was a nuisance, so the machine’s status lights were wired up so that people could check from their desks which columns had stock and whether recent bottles were likely to be cold. It was a small departmental project rather than a product, but it is often cited as an early example of an everyday object reporting its state over a network.
A similar story from 1991 involved a coffee pot in a computer laboratory at the University of Cambridge. A camera pointed at the pot sent pictures over the local network so researchers could see whether coffee was left before making the trip. The images were later put on the web, and the pot became an early well-known webcam.
Computing everywhere
In 1991 Mark Weiser of Xerox PARC wrote about “ubiquitous computing,” imagining that computers would fade into the background of everyday life, embedded in walls, desks and tools. This was more a vision than a working system, but it gave researchers a shared direction. Devices such as programmable thermostats and early home automation products showed that simple electronics could be placed in ordinary household items.
Radio tags and supply chains
The second root of the idea came from logistics. Radio-frequency identification (RFID) tags are small chips that reply with an identifying number when read by a nearby device. Variations of the technology were used in the mid-twentieth century to identify aircraft, and later in applications such as livestock tracking and toll collection. Compared with a barcode, a tag does not have to be seen by a scanner, and many can be read together.
In the late 1990s, a group at the Massachusetts Institute of Technology, the Auto-ID Center, worked on cheap tags and a shared network for looking up information about tagged items. The aim was a system in which a pallet or product could be identified anywhere in a supply chain without a person scanning it.
The term itself
The phrase “Internet of Things” is commonly credited to Kevin Ashton, who used it in 1999 in a presentation about linking RFID in supply chains to the internet while working with Procter & Gamble and the Auto-ID Center. His point was that people had been the main source of information on the internet, typing or photographing things, and that computers able to gather data about the physical world themselves would be more accurate and less costly. The phrase caught on gradually through the 2000s.
Why it became practical
Several trends made the idea feasible for ordinary products:
- Small, low-power chips and sensors became cheap, partly through ongoing integrated circuits miniaturization.
- Wireless links such as Wi-Fi, Bluetooth and cellular networks spread widely, so devices no longer needed cables.
- Smartphones supplied a convenient screen and control for gadgets that had none of their own.
- Cloud computing provided inexpensive places to store and process what devices reported.
By the 2010s, products such as connected thermostats, doorbells with cameras, fitness trackers, smart speakers and meters that reported usage to utilities had become familiar. Factories and farms also adopted connected sensors to monitor machines, soil or livestock. Estimates of how many devices are online vary widely depending on definition, so it is safest to say simply that there are billions.
Problems that came with it
Many early connected products were built with little attention to security, and poorly protected devices have been taken over and used in attacks, an issue now central to cybersecurity. Devices can also collect personal information about habits at home. Some products stop working when a manufacturer ends support for its online service, leaving working hardware useless. Competing standards mean that gadgets from different makers do not always cooperate.
The history shows that the Internet of Things was not one invention but a convergence of vending-machine hacks, tagging technology and falling hardware costs. How much everyday life will ultimately depend on it remains to be seen.