The Internet of Things (IoT) is the idea of extending network connectivity beyond computers and phones to ordinary physical objects: thermostats, light bulbs, door locks, industrial sensors, shipping pallets, and machinery. Each object carries some combination of sensors, a small processor, and a radio, and reports its state or receives commands over a network. The goal is to let computers observe and influence the physical world directly, instead of waiting for people to type the data in.
The phrase itself has a documented origin. In his 2009 RFID Journal article “That ‘Internet of Things’ Thing,” Kevin Ashton wrote, “I could be wrong, but I’m fairly sure the phrase ‘Internet of Things’ started life as the title of a presentation I made at Procter and Gamble (P&G) in 1999.” Ashton’s point was that nearly all data on the internet of that era was created by humans through typing, recording, or scanning, and that people have “limited time, attention and accuracy.” Computers, he argued, needed their own means of gathering information about the physical world through technologies like RFID and sensors.
Architecturally, IoT systems are usually layered. Constrained end devices (often built on cheap microcontrollers) sense or actuate; they talk over short-range or low-power radios to a gateway; the gateway bridges to the wider internet, where data flows to cloud services for storage and analysis. This structure is reflected in the protocols built for it. RFC 7252, the Constrained Application Protocol, describes its target devices as “8-bit microcontrollers with small amounts of ROM and RAM” on networks with “high packet error rates and a typical throughput of 10s of kbit/s,” a vivid picture of how different IoT endpoints are from ordinary web servers.
The scale is what makes IoT distinctive. Forecasts routinely speak of tens of billions of connected devices, far outnumbering the world’s people. That scale forces design choices at every level: protocols must be lightweight, radios must sip power so a battery lasts for years, and management must be automatic because no one will manually configure billions of gadgets.
These constraints gave rise to a family of purpose-built standards. Lightweight application protocols such as CoAP and MQTT replace heavyweight HTTP for messaging; mesh standards such as Zigbee and long-range standards such as LoRaWAN replace Wi-Fi for the radio link; and edge computing pushes processing closer to the devices to cut latency and bandwidth. Together they form the plumbing that the marketing term “Internet of Things” describes.
The same properties that make IoT useful also make it dangerous. Devices that are cheap, numerous, long-lived, and rarely updated are hard to secure, and the consequences of that fragility have been demonstrated at internet scale, a story told in the entry on the IoT security problem.