The internet of things (IoT) is a catch-all term for the growing number of electronics that aren't traditional computing devices, but are connected to the internet to send data, receive instructions or both.\nThere's an incredibly broad range of \u2018things\u2019 that fall under the IoT umbrella: Internet-connected \u2018smart\u2019 versions of traditional appliances such as refrigerators and light bulbs; gadgets that could only exist in an internet-enabled world such as Alexa-style digital assistants; and internet-enabled sensors that are transforming factories, healthcare, transportation, distribution centers and farms.\nWhat is the internet of things?\nThe IoT brings internet connectivity, data processing and analytics to the world of physical objects. For consumers, this means interacting with the global information network without the intermediary of a keyboard and screen (Alexa, for example).\nIn enterprise settings, IoT can bring the same efficiencies to manufacturing processes and distribution systems that the internet has long delivered to knowledge work. Billions of embedded internet-enabled sensors worldwide provide an incredibly rich set of data that companies can use to improve the safety of their operations, track assets and reduce manual processes.\nData from machines can be used to predict whether equipment will break down, giving manufacturers advance warning to prevent long stretches of downtime. Researchers can also use IoT devices to gather data about customer preferences and behavior, though that can have serious implications for privacy and security.\nHow big is the IoT?\nIn a word: enormous. Priceonomics breaks it down: There were more than 50 billion IoT devices in 2020, and those devices generated 4.4 zettabytes of data. (A zettabyte is a trillion gigabytes.) By comparison, in 2013 IoT devices generated a mere 100 billion gigabytes. The amount of money to be made in the IoT market is similarly staggering; estimates on the value of the market in 2025 range from $1.6 trillion to $14.4 trillion.\nIn its Global IoT Market Forecast, IoT Analytics Research predicts there will be 27 billion active IoT connections (excluding computers, laptops, phones, cellphones and tablets) by 2025. However, the company did lower its forecast based on the ongoing chip shortage, which it expects to impact the number of connected IoT devices beyond 2023.\nHow does the IoT work?\nThe first element of an IoT system is the device that gathers data. Broadly speaking, these are internet-connected devices, so they each have an IP address. They range in complexity from autonomous mobile robots and forklifts that move products around factory floors and warehouses, to simple sensors that monitor the temperature or scan for gas leaks in buildings.\nThey also include personal devices such as fitness trackers that monitor the number of steps individuals take each day.\nIn the next step in the IoT process, collected data is transmitted from the devices to a gathering point. Moving the data can be done wirelessly using a range of technologies or over wired networks. Data can be sent over the internet to a data center or the cloud. Or the transfer can be performed in phases, with intermediary devices aggregating the data, formatting it, filtering it, discarding irrelevant or duplicative data, then sending the important data along for further analysis.\nThe final step, data processing and analytics, can take place in data centers or the cloud, but sometimes that\u2019s not an option. In the case of critical devices such as shutoffs in industrial settings, the delay of sending data from the device to a remote data center is too great. The round-trip time for sending data, processing it, analyzing it and returning instructions (close that valve before the pipes burst) can take too long.\nIn such cases edge computing can come into play, where a smart edge device can aggregate data, analyze it and fashion responses if necessary, all within relatively close physical distance, thereby reducing delay. Edge devices also have upstream connectivity for sending data to be further processed and stored.\nA growing number of edge computing use cases, such as autonomous vehicles that need to make split-second decisions, is accelerating the development of edge technologies that can process and analyze data immediately without going to the cloud.\n\n Network World \/ IDG\n\nHow the internet of things works.\u00a0\n\n\nExamples of IoT devices\nEssentially, any device that can gather and transmit information about the physical world can participate in the IoT ecosystem. Smart home appliances, RFID tags, and industrial sensors are a few examples. These sensors can monitor a range of factors including temperature and pressure in industrial systems, status of critical parts in machinery, patient vital signs, the use of water and electricity, among many, many other possibilities.\nFactory robots can be considered IoT devices, as well as autonomous vehicles and robots that move products around industrial settings and warehouses. Municipalities exploring smart city ecosystems are using IoT and machine-to-machine (M2M) sensors to enable applications such as traffic monitoring, street light management, and crime prevention through camera feeds.\nOther examples include fitness wearables and home security systems. There are also more generic devices, like the Raspberry Pi or Arduino, that let you build your own IoT endpoints. Even though you might think of your smartphone as a pocket-sized computer, it may well also be beaming data about your location and behavior to back-end services in very IoT-like ways.\nIoT device management\nIn order to work together, all those devices need to be authenticated, provisioned, configured, and monitored, as well as patched and updated as necessary. Too often, all this happens within the context of a single vendor's proprietary systems \u2013 or, it doesn't happen at all, which is even more risky. But the industry is starting to transition to a standards-based device management model, which allows IoT devices to interoperate and will ensure that devices aren't orphaned.\nIoT communication standards and protocols\nWhen IoT gadgets talk to other devices, they can use a wide variety of communication standards and protocols, many tailored to devices with limited processing capabilities or low power consumption. Some of these you've definitely heard of \u2014 Wi-Fi or Bluetooth, for instance \u2014 but many more are specialized for the world of IoT. ZigBee, for example, is a wireless protocol for low-power, short-distance communication, while message queuing telemetry transport (MQTT) is a publish\/subscribe messaging protocol for devices connected by unreliable or delay-prone networks. (See Network World\u2019s glossary of IoT standards and protocols.)\nThe increased speeds and bandwidth of 5G cellular networks are expected to benefit IoT. In its Global IoT Market Forecast, IoT Analytics Research predicted a compounded annual growth rate (CAGR) of 159% for 5G-based IoT devices from 2021 through 2025.\nIoT, edge computing and the cloud\n\u00a0\n Network World \/ IDG\n\nHow edge computing enables IoT.\n\n\nFor many IoT systems, the stream of data is coming in fast and furious, which has given rise to a new technology category called edge computing, which consists of appliances placed relatively close to IoT devices, fielding the flow of data from them. These machines process that data and send only relevant material back to a more centralized system for analysis. For instance, imagine a network of dozens of IoT security cameras. Instead of bombarding the building's security operations center (SoC) with simultaneous live-streams, edge-computing systems can analyze the incoming video and only alert the SoC when one of the cameras detects movement.\nAnd where does that data go once it\u2019s been processed? Well, it might go to your centralized data center, but more often than not it will end up in the cloud. The elastic nature of cloud computing is great for IoT scenarios where data might come in intermittently or asynchronously.\nCloud vendors offer IoT platforms\nThe cloud giants (Microsoft, Amazon, Google) are trying to sell more than just a place to stash the data your sensors have collected. They're offering full IoT platforms, which bundle together much of the functionality to coordinate the elements that make up IoT systems. In essence, an IoT platform serves as middleware that connects the IoT devices and edge gateways with the applications you use to deal with the IoT data. That said, every platform vendor seems to have a slightly different definition of what an IoT platform is, the better to distance themselves from the competition.\nIoT and Big Data analytics\nImagine a scenario where people at a theme park are encouraged to download an app that offers information about the park. At the same time, the app sends GPS signals back to the park's management to help predict wait times in lines. With that information, the park can take action in the short term (by adding more staff to increase the capacity of some attractions, for instance) and the long term (by learning which rides are the most and least popular at the park).\nThe theme park example is small potatoes compared to many real-world IoT data-harvesting operations. Many big data operations use information harvested from IoT devices, correlated with other data points, to get insight into human behavior.\nFor example, X-Mode released a map based on tracking location data of people who partied at spring break in Ft. Lauderdale in March of 2020, even as the coronavirus pandemic was gaining speed in the United States, showing where all those people ended up across the country. The map was shocking not only because it showed the potential spread of the virus, but also because it illustrated just how closely IoT devices can track us. (For more on IoT and analytics, click here.)\nIoT and AI\nThe volume of data IoT devices can gather is far larger than any human can deal with in a useful way, and certainly not in real time. We've already seen that edge computing devices are needed just to make sense of the raw data coming in from the IoT endpoints. There's also the need to detect and deal with data that might be just plain wrong.\nMany IoT providers are offering machine learning and artificial intelligence capabilities to make sense of the collected data. IBM's Watson platform, for instance, can be trained on IoT data sets to produce useful results in the field of predictive maintenance \u2014 analyzing data from drones to distinguish between trivial damage to a bridge and cracks that need attention, for instance. Meanwhile, Arm has announced low-power chips that can provide AI capabilities on the IoT endpoints themselves. The company also launched new IoT processors, such as the Cortex-M85 and Corstone-1000 that supports AI at the edge.\nIoT and business applications\nBusiness uses for IoT include keeping track of customers, inventory, and the status of important components. Here are four industries that have been transformed by IoT:\n\nOil and gas: Isolated drilling sites can be better monitored with IoT sensors than by human intervention.\nAgriculture: Granular data about crops growing in fields derived from IoT sensors can be used to increase yields.\nHVAC: Climate control systems across the country can be monitored by manufacturers.\nBrick-and-mortar retail: Customers can be micro-targeted with offers on their phones as they linger in certain parts of a store.\n\nMore generally, enterprises are looking for IoT solutions that can help in four areas: energy use, asset tracking, security, and customer experience.\nIndustrial IoT\nThe IIoT is a subset of the Internet of Things made up of connected sensors and instrumentation for machinery in the transport, energy, and industrial sectors. The IIoT includes some of the most well-established sectors of the IoT market, including the descendants of some devices that predate the IoT moniker. IIoT devices are often longer-lived than most IoT endpoints \u2013 some remain in service for a decade or more \u2013 and as a result may use legacy, proprietary protocols and standards that make it difficult to move to modern platforms.\nConsumer IoT\nThe move of IoT into consumer devices is more recent but much more visible to ordinary people. Connected devices range from fitness wearables that track our movements to internet-enabled thermometers. Probably the most prominent IoT consumer product is the home assistant, such as Amazon Alexa or Google Home.\nIoT security and vulnerabilities\nIoT devices have earned a bad reputation when it comes to security. PCs and smartphones are "general use" computers designed to last for years, with complex, user-friendly OSes that now have automated patching and security features built in.\nIoT devices, by contrast, are often basic gadgets with stripped-down OSes. They are designed for individual tasks and minimal human interaction, and cannot be patched, monitored or updated. Because many IoT devices are ultimately running a version of Linux under the hood with various network ports available, they make tempting targets for hackers.\nPerhaps nothing demonstrated this more than the Mirai botnet, which was created by a teenager telnetting into home security cameras and baby monitors that had easy-to-guess default passwords, and which ended up launching one of history's largest DDoS attacks.\nWhile this situation is improving somewhat, the truth is that IoT transactions are by and large still not secure. Enterprise IoT customers can work to improve their IoT security, but vendors need to make IoT devices more secure and easier to keep secure if they're going to be in the field for any length of time. (Related story: 7 steps to enhance IoT security)\nCompanies continue to expand on their security options for IoT devices. For example, Microsoft recently announced its Edge Secured-core program for Windows-based IoT devices. The program aims to address issues such as device identity, secure boot, operating system hardening, device updates, data protection, and vulnerability disclosures.\nPrivacy and IoT\nNo one wants a hacker snooping on private data. But what if it's the company that sold the gadget that's doing the spying? Take, for instance, home digital assistants. Recode has a pretty good breakdown on what Amazon and Google can learn about a person from connected devices. You're probably not that worried about Amazon learning when you turn your lights on and off, but remember, every bit of information goes into a data lake that can help companies produce a surprisingly complete picture of your life.\nThe ability of IoT devices to track a user's location is a particular privacy concern. Consider the map of spring breakers potentially carrying coronavirus home, for instance. Location data is in theory anonymized, but The New York Times put together an extensive report showing how, for instance, the data could be used to track the movements of individuals who could later be identified by name. Another incident crossed the line into a failure in operational security: the Strava heat map, which shows popular running routes for Fitbit users around the world, accidentally revealed several secret American military bases.\nHistory of IoT\nA world of omnipresent connected devices and sensors is one of the oldest tropes of science fiction. IoT lore has dubbed a vending machine at Carnegie Mellon University that was connected to ARPANET in 1970 as the first Internet of Things device, and many technologies have been touted as enabling "smart" IoT-style characteristics to give them a futuristic sheen. But the term Internet of Things was coined in 1999 by British technologist Kevin Ashton.\nAt first, the technology lagged behind the vision. Every internet-connected thing needed a processor and a means to communicate with other things, preferably wirelessly, and those factors imposed costs and power requirements that made widespread IoT rollouts impractical, at least until Moore's Law caught up in the mid-2000s.\nOne important milestone was widespread adoption of RFID tags, cheap minimalist transponders that can stick to any object to connect it to the larger internet world. Omnipresent Wi-Fi, 4G and 5G wireless networks make it possible for designers to simply assume wireless connectivity anywhere. And the rollout of IPv6 means that connecting billions of gadgets to the internet won't exhaust the store of IP addresses, which was a real concern. (Related story: Can IoT networking drive adoption of IPv6?)\nWhat's next for IoT?\nAs the number of IoT devices continue to grow, companies will continue to improve security features and look to faster connectivity options, such as 5G and faster Wi-Fi, to enable more functionality for getting the data processed and analyzed. Additional collaboration between IT and operational technology (OT) is also expected.\nIoT will continue to grow as smaller companies get in on the action, and larger enterprises and industry giants such as Google and Amazon continue to embrace IoT infrastructures. Perhaps IoT devices will become so omnipresent that a wireless network can essentially be treated as a giant sensor. The technology continues to be exciting to watch.