Why The Internet of Things Is Going Nowhere
The next phase of the IoT is stuck unless we replace crummy outdated technology
How The IoT Was Blindsided
Once upon a time, some people in the Defense Department created the internet. Then when that turned out pretty well, some other people decided to try to create a wireless version of the internet and invented technologies like WiFi. Still others invented more ways of connecting gadgets to this internet thing and built technologies like Bluetooth and ZigBee. All of this happened a long time ago before MIT even coined the term “internet of things”.
But a funny thing happened on the way to the internet of things: newer and more exciting technologies. ARM-based processors, MEMS sensors and actuators, the iPhone, and more. It all happened so fast … and as I’ve outlined here and here and here, today’s wireless IoT technologies are the cassette tapes of the IoT and won’t scale in any kind of secure or practical way. But today’s wireless IoT pain looks like pleasure compared to the ugly reality of the mobile IoT, or what I like to call the “Internet of Moving Things”.
Why Movement is Important to the IoT
An Internet of Moving Things is at its core a network of physical objects that are mobile or moveable and can be wirelessly measured or controlled. Amazingly, the designers of today’s IoT gave little or no thought to connecting things that move and today we are stuck with wireless IoT technologies that require a lengthy pairing ritual that renders them nearly useless for connecting with a moving thing. Yet solving for moving things is of critical importance to the future of the IoT because:
- Things move! An IoT that cannot measure or control things that move or things that are moving is an IoT that … is an Internet of Only Some Of The Things. It’s like the internet without SMTP email.
- Movement = news. When a thing moves, it is often an important or “newsworthy” event. For example, sensing movement on your parked bicycle might be an indication that it is being stolen.
- Location. Moving things often change geographic location. The question “Where is it?” is asked every day by billions of people about everything from car keys to shovels to humvees. DoD spends millions of hours every year just looking for stuff like humvees, fyi.
- Cause and Effect. When things move, it’s often caused by or associated with someone or something else. “Movers”, as Aristotle might say, are highly relevant to the big data questions being asked every day. “Who was the last guy to use this broken chainsaw?” or “Which truck carried these frozen vials of flu vaccine?”
- Absence. Conversely, an absence of movement can itself be meaningful. A cow on a ranch or a stay-at-home senior that does not move for 24 hours might be injured or sick.
- Battery life. Movement can be an important power saving tool for managing sleep/wake cycles battery-powered devices. A mobile device that doesn’t move might sleep more than one constantly on the move.
- Movement changes risk. A moving thing may set in motion other events — or at least the potential for other events to occur. A moving car or bicycle or fighter jet is far more likely to crash or experience mechanical failure than one that is parked.
Moving Things Are Already Making Today’s IoT Obsolete
While the mobile internet is expected to be 10x the size of the desktop internet, the forecasts seem not to have affected most of the folks working on the IoT.
For example, drones are forecasted to be a multi-billion dollar industry unto themselves and are already showing their potential as “IoT gateways in the sky” for everything from monitoring oil pipelines to moisture content on farmland. But the go-to wireless technologies for today’s drones are … wait for it … WiFi and Bluetooth, which can’t communicate with terrestrial-based sensors while moving because they were designed 20+ years ago when CompuServe was still hot.
Along with Bluetooth, ZigBee, Thread, and others, WiFi radios perform an elaborate mating ritual when they connect to a new endpoint. If you’ve ever logged onto a new WiFi access point and it took many seconds or even minutes, it’s because there’s an outdated and obsolete sequence of discovery, handshaking, authentication, payload delivery, and more. As a result, to connect with any of these, you must remain still — completely still, usually — in order to establish a connection.
For local area networks like your house where endpoints and access points are mostly fixed, this is probably not an issue. But for things that move at speeds of 5 mph or more, requiring endpoints to stop whatever they are doing and remain still while a wireless technology “locks on” is simply a non-starter.
Note to cellular people reading this: you can drive a drone from afar using cellular, but expecting the rest of the IoT to deploy high-cost and high-powered cellular at the endpoint is fantasy apart from some niche use cases where cost and battery life are not important.
A Fast Moving World Needs an IoT That Can Keep Up
Drones are just one example of a world of moving things challenging the capabilities of today’s IoT. Here are roughly four categories of things to solve for:
1. Mobile Vehicles: Next-Gen IoT Gateways
Mobile vehicles have important roles to play as messaging and sensor data gateways that stand between endpoints and cloud-based applications. Most have onboard power supplies (e.g. a car battery) and most will have high speed cellular connections that facilitate this gateway role. The ability for a fast moving mobile IoT gateway to reliably communicate with fixed (“vehicle-to-infrastructure”) endpoints is a significant gap in the capabilities of today’s IoT.
Mobile IoT gateways will also communicate with other mobile IoT gateways. For example, there is already work underway around vehicle-to-vehicle or “V2V” wireless using variations on short-range WiFi to support collision avoidance, but this solves (in a very non-secure way) for only part of the vehicle-based opportunity which includes not only other vehicles, but also mobile and fixed IoT endpoints placed along highways, bridges, city streets, or parking garages.
2. IoT Endpoints That Move
These are the millions or billions of endpoints that move, spanning a array of industrial, commercial, government, and consumer applications. When solving for moving IoT endpoints, understanding the speed (or distance traveled per second) of an object is important to arriving at the right connectivity solution. For example, there are things that move pretty slowly:
And there are the things that can move fast:
There are things that don’t move on their own but get moved around:
3. IoT Endpoints That Don’t Move
A large part of an Internet of Moving Things is just using mobile IoT gateways to measure fixed things. These cover a diverse array of applications and industries and many will be without mains power and require a battery, which in turn means the endpoint can’t use a power-thirsty wireless technology lest it require frequent battery changes or recharges — the near-universal achilles heel of IoT endpoints.
4. Smartphone Gateways and Endpoints
The smartphone represents an exciting opportunity for mobile IoT data acquisition as I’ve outlined here and clearly smartphones are their own class of mobile endpoint. Smartphones should be center stage in the IoT but without better wireless IoT connectivity, we’ll just have to enjoy our crummy IoT options via WiFi and Bluetooth until you-know-who decides to take the lead.
Four Killer Apps for the Internet Of Moving Things
I meet many developers with cool IoT ideas and here are a few that I consider to be killer apps for an Internet of Moving Things:
Driverless cars. To be successful, driverless cars are going to need help from wireless IoT technologies to solve some big, unsolved vehicle-to-infrastructure problems including detecting icy bridges, pedestrians, potholes, uncovered manholes, operating in snow and rain, or maneuvering around construction road crews. Interfacing securely and in real-time with other moving or parked cars more than a few feet away is an unsolved problem as well as communicating with “smart city” and other battery-powered fixed infrastructure like smart street signs or bridge stress sensors. Solving for V2I (and vehicle-to-pedestrian, “V2P”) is fundamentally a low power IoT problem (most V2I situations will not include access to mains power sources) and while the auto industry’s current dalliance with WiFi may make for some nice demos, it won’t solve for V2I or V2P in any secure, private, safe, or scalable way.
Agriculture Drones. Some of the most exciting drone apps focus on the agricultural sector where drones not only provide high-resolution imagery and infrared sensing, but also an “IoT gateway in the sky” capturing data from animals, water troughs, tools and equipment in the field, and more. The opportunity extends just as easily to other industries where drones are becoming popular like oil and gas, mining, construction, and defense.
Search and Rescue. Locating missing firefighters or hikers in a forest, shipwreck survivors, lost dogs, and more via moving vehicles like drones or cars. The “eyeball” method for these use cases is now officially obsolete and fast moving vehicles need a wireless technology that can locate things in hard to reach, hard to see places.
E-commerce. A killer mobile IoT opportunity — perhaps with a longer time horizon? — is in the area of e-commerce, where everything from the handbag carried by a woman walking her dog to passing buses will effectively become a query-able, wireless IoT endpoint. Soon, we will be able to learn the make and model of a cool mountain bike as it passes by or capture a promo code as we pass a billboard on a highway.
How To Get The IoT Moving
The big question to explore: what is the best wireless connectivity option to address the internet of moving things? Here are the must-have requirements based on everything we’ve learned so far:
- Instant-on. Solving for the Internet of Moving Things absolutely requires the ability to have “instant on” connectivity. A basic scenario of a fixed endpoint connecting with a car or truck moving at 60 mph means there might be 1–2 seconds of time available to send and receive a message before the vehicle is out of range. This is closely related to the concept of real-time queries at the endpoint which is explored in more detail here.
- Cross-Application Data Interoperability. On our PCs, we know about the difference of file formats, but we take for granted the fact that there is a common concept of a file, containing a name, modified date, and other useful information along with the data. There is no such standardization with IoT platforms. The disconnect, here, occurs because most standards bodies are too compartmentalized: teams work on the wireless standardization, others the networking, others the applications. In an IoT of moving things, it is excessively important for data exchange to be transactable between devices owned by different owners, managed by different networks, and sending different kinds of data. The concept of the file must be clear, and it must be fundamental to the technology stack.
- P2P. Similar to instant-on, capturing data from a fixed or moving object — like a driverless car — should not require a time-consuming cloud lookup and/or cellular network connection and should instead provide the option of pure point-to-point connectivity. Bluetooth Low Energy, for example, operates on the principle of one-way “beacons” that provide a unique endpoint identifier that can only be resolved via a cloud lookup. As hacks go, Bluetooth has done about the best it can do but the IoT cannot be built around a hack. P2P is also important from the perspective of associating an event with other nearby endpoints. E.g., “The door to the refrigerator containing the possible cancer cure was left open by Roger at 2:13 a.m. last night.”
- Plays Well with Thousands of Endpoints. Crowded environments like smart cities and dense industrial settings will be overflowing with fixed and mobile IoT endpoints. Using non-stop beaconing for these is a non-starter, lest we run out of available radio spectrum, drain endpoint batteries, and violate basic principles of privacy. It must be possible to quickly perform queries from among thousands or even millions of endpoints in a given area and with no degradation of latency.
- Good signal range. The farther your wireless IoT technology can communicate, the better its chances of connecting with a moving object, especially one moving at 60+ mph. For example, the short range of technologies like Bluetooth (30 feet max, often less) make connecting with moving objects next to impossible. In 2015 there is no reason not to demand maximum range measured in miles.
- Stealth. It’s one thing for your home thermostat to be hacked but it’s quite another thing for your car’s accelerator or your child’s location to be hacked. “Stealthy” endpoints that listen before they talk (explored in greater detail here) with better low level (MAC layer) encryption are minimally required versus the comically bad privacy and security found in Bluetooth, WiFi, ZigBee, and other technologies derived from Ethernet. The non-stop beaconing hack used in some V2I demos is one example of security comedy that could end tragically, but anyone thinking of using variants of a notoriously insecure protocol like WiFi for should remember what happens to automakers who cut corners.
- Low Power. Since most IoT endpoints, either fixed or mobile, are battery powered, low power remains a must-have for nearly all devices in the Internet of Moving Things. For niche devices or mobile IoT gateways (like a smartphone) that customers are already resigned to recharging every day or two (cellular technology is notoriously high power), this requirement does not apply. For 99% of other IoT endpoints, multi-year battery life is non-negotiable.
- Low Cost. Like low power, the non-ascendancy of cellular as a mass market IoT technology is also attributed to the high costs of cellular chipsets and the recurring monthly fees from cellular carriers working to amortize the high capital costs of running a cellular network. Suggested retail price for battery powered IoT modules should be less than $10 based on my own discussions with hundreds of developers and customers, many of whom have a target price point well below $10.
Preparing Your Company For The Internet of Moving Things
If your company is throwing elbows and punches inside the mosh pit of home automation, perhaps the Internet of Moving Things is not your biggest priority. However if your company is investing in smart cities, driverless cars, drones, wearables, or any of the other industries mentioned above, you should be taking note of the shortcomings of the most common wireless IoT technologies in the marketplace as it relates to movement. In some cases it may be possible to demand incumbents update their technologies to better support movement, in other cases the right decision will be to start afresh. My company, Haystack, is one of the few in the IoT marketplace that designed its IoT networking stack (based on DASH7) with the Internet of Moving Things in mind from the beginning.
But one thing is certain: a failure to address movement will bode poorly not ony for the IoT, but for many new innovations, including many we haven’t even imagined, that will depend on an IoT that can adequately communicate with things that move. The same way that “unforeseen” companies like Uber took advantage of foundational breakthroughs like smartphones, Google Maps, and low cost GPS chipsets, similar companies will be built around an IoT that properly addresses movement.
Get the latest about Haystack by signing up for our newsletter here.