The internet of things combined with cloud computing is the platform for innovation that is used by NASA’s Jet Propulsion Laboratory and that should be used by enterprises, but it means setting up the right network infrastructure, JPL’s IT CTO says.
“Number one, build an IoT network that’s separate from the regular network,” says Tom Soderstrom, the JPL IT CTO. “That’s what we did, and we found that it was amazing.”
That separation is important because it eliminates the cybersecurity threats that lurk on the open internet. When it is needed, secure bridges can be made to reach the internet, he says. “We built code, programs, that talk back and forth between the two.”
The goal for JPL and for enterprises should be to experiment freely with inexpensive consumer IoT devices to discover whether they might have practical business uses, he says.
For example, sensors are a large part of IoT, and each industry has its own type. “Imagine in the shipping industry, each container tells its health, its temperature, its everything,” Soderstrom says.
Another example: “We took a toy [Mars] Rover that you can talk to and drive and you can ask questions about Mars, and it drives around and follows you,” he says. That can be a model for real-world applications.
He says in the mining industry, a similar robot could descend into a mine to look for trapped miners, then reemerge and connect to the internet to disseminate what it found and to receive new instructions.
“Imagine that in the trucking industry…a [driverless] truck driving vast distances across Australia where there’s no internet but it still keeps going. When it gets in range, it connects back up again.”
The low cost of IoT devices combined with the low cost of cloud computing means enterprises can vastly expand the trials they run, and experimental failure holds fewer negative consequences. “Take this new technology and apply it into the enterprise. Fast. High-risk,” he says. “Reward failure, because failure doesn’t cost anything in this case.”
He describes the aggressive experimental teams working on these projects by borrowing a term he picked up from Amazon – the pizza team, meaning a group working on a project together that is small enough to share a pizza.
“So every new technology that comes in, we execute it in the cloud in a one-pizza team or a two-pizza team, small groups that can experiment. If it doesn’t work, that’s OK. We didn’t pay for it, very much. And then we can iterate much, much, much faster. We live in a time of accelerating exponential curves, and this is one way to take advantage of it,” he says.
Sign up for CIO Asia eNewsletters.