Wednesday, October 28, 2015

LPWAN for M2M

The phrase "the Internet of things" just irritates me. I don't want my washing machine to wirelessly connect to my laptop. I don't need my TV to report data to my printer, cellphone or my sump pump. I don't need to wirelessly cam my cat while I'm at work, and my hot water heater and my deep freezer do not need to be friends. I don't need a personal drone to deliver used books to my doorstep. I can wait a few days for USPS. I even find the relatively inoffensive Fitbit somewhat dubious. But as a radio geek you knwo I'm no Luddite. I just have no use for bells and whistles that increase the cost of goods. Ex. car back up cameras. I have three mirrors and a neck. I can just turn and look. No thank you.

But it's coming anyway. The obstacles to the great wireless everything have mostly been in the realm of physics: battery life and signal distance.  Then there is one legal one: signal strength. If your output exceeds a certain amplitude, which varies by band, then your device requires a nod from the FCC. All three of these obstacles have been conquered in the last year by a set of technologies collectively called LPWAN - Low-Power Wide Area Networks. It's one answer for what's called M2M, Machine-to-Machine communications. (M2M is also a terrible Norwegian pop duo) There is a lot of money on the table now, so the VC articles are coming out in force: here, here, and here.

Bluetooth for example is very power efficient, but only has about a 30ft range because the output power is very low, 3 watts at the maximum. A cell tower can operates at 1000 watts but your cell phone operates at about 3 watts but consumes much more power as you can tell by how often you have to charge it. (Mine is daily but that's diminished by MDM software.) Zigbee (the open, global wireless standard) is only good to about 60 feet.  The scale of this technical problem grows with the number of devices we try to connect. How big will that be?  In 2012 Cisco has claimed there would be 50 billion devices online by 2020.  Brett Swanson at Computerworld echoed that comment just a month ago:
"As the world becomes saturated with 7 billion or so mobile phones, however, the immersive Internet will multiply that total several times over. In the next decade, we might connect 50 billion or so devices to the Internet — every car, watch, boat, shoe, package, shipping container, medical device, vending machine, camera, drone, you name it."

So radio engineers have been hard at work on a solution. Computer World even recently referred to the existing technologies as "unapologetically slow." Numerous brands have worked out their own solutions: LoRaWAN, Semtech, Sigfox, and Ingenu. The goal is not to make the data move faster but to make the range further and the technology cheaper. It doesn't matter how LPWAN gets there and there are already multiple solutions on the market that all use unlicensed wireless spectrum burned into a silicon chip. It's that affordability that will make it ubiquitous.