Trillions and trillions of devices everywhere

Somewhere in my font of near-useless TV trivia sticks a moment that perfectly defines marketing the Internet of Things. Searching IMDb quickly revealed the source of the memory as “Shadow Chasers”, a short-lived 1985 TV series about a team searching for the paranormal.

In the opening minutes of the first episode, we meet tabloid reporter Edgar ‘Benny’ Benedek (portrayed by Dennis Dugan of “Happy Gilmore” fame) who has a high-concept pitch for a story: Elvis Presley as extraterrestrial. His editor says he needs some kind of proof to run with it, and he needs it fast to make deadline. Benny thinks for a few seconds, and then picks up the phone.

Carl Sagan

He calls Carl Sagan, astronomer and author and host of “Cosmos”, during the 1980s the most widely watched series in the history of public television. We only hear Benny’s side of the conversation asking Sagan point-blank about the possibility that Elvis was an alien, and verifying Sagan’s response saying he was astonished that anyone would call him and ask such a ludicrous question. The resulting headline emblazoned across the next edition:

Elvis Was An Alien, Sagan Astonished

I’m reminded of that every time I see the beloved “hockey stick”. Hyperbolic technology growth projections extrapolated from a shred of fact and voiced with conviction by industry leaders draw in more participants. Take this headline, inspired by the recent TSensors Summit:

Millions are boring, billions are interesting, but trillions get people all worked up if the story is good enough. This assertion is completely plausible even if actual growth doesn’t exactly match the hyperbole, as charted below in various projections of worldwide sensor population.

T Sensors forecast

Somewhere between the popular figure of 50 billion devices and this new discussion of trillions of sensors is the IoT event horizon. The idea behind the summit was to look at some of the issues in developing a roadmap for the required technology improvements. Besides the link in the Tweet above, another good summary of the summit highlights is in Rick Merritt’s EDN article.

This was a sensors summit, and many of the issues they identified were specific to MEMS and RFID components. In order for sensors to become part of IoT devices, a few other items are needed.

More microcontrollers: At the heart of many billions of devices will be the microcontroller, and we will need a lot more to get to a trillion units. According to IC Insights, we are on a run rate of give-or-take 25 billion MCUs of all shapes and sizes every year. Many of these are still the 4-bit or 8-bit variety, with limited processing power for small tasks – presumably, most IoT devices will feature 16- and 32-bit MCUs. With ARM well established, MIPS still a contender with the backing of Imagination Technologies, and the entry of Intel and their Quark processor core brings something badly needed: sheer manufacturing capacity.

MCU forecast

More energy: Trillions of devices, even advanced ultra-low power ones, require a lot more energy. As Freescale CEO Gregg Lowe pointed out at DAC 2013, all those connected devices bring along networking, server, and storage resources in data centers, either centralized or in the “cloud”. The irony is individual devices may be deployed to save power, but the aggregate weight of trillions of devices will tax the limits of data centers, already the largest industrial consumers of electricity.

More joules: Most of these devices are no fun plugged in, and need either batteries or self-powering technology – and both need vast improvements. Thin film batteries like those from Cymbet and Imprint Energy are just a start. Energy harvesting components, particularly those using piezoelectric ceramics that can pull electricity from vibration, need to improve in output by an order of magnitude and also solve degradation issues.

More connectivity: Becoming a connected device requires a wireless sensor network (WSN), and we are starting to see the needed maturity in deployments. Bluetooth Low Energy is gaining support in every smartphone and tablet. Gateways and hubs are starting to show up with combinations of WSNs for home automation, such as the Revolv with support for INSTEON, Wi-Fi, Z-Wave, and soon ZigBee.

Revolv Home Automation Hub

More software: Two dimensions of improvement are needed. Web services programming using HTML is effective for bigger computers, but for smaller devices something like CoAP or MQTT (interestingly, a technology found in Facebook) will likely gain popularity. Data interchange formats are also needed for “big data” and interoperability; beyond simple JSON objects, networks will need interchange frameworks such as the IEEE 11073 standard for healthcare information.

A trillion is a lot of anything, but the vision that sensors and IoT devices can be the biggest business in the history of electronics carries a lot of credibility. What do you think of the ideas raised in the TSensors Summit, or the above suggestions? What other technological improvements do you see needed to drive IoT adoption?

, , , , , , , , , , ,