Subscribe to the O’Reilly Radar Podcast to track the technologies and people that will shape our world in the years to come.
In this week’s Radar Podcast, I followed up with Eric Jennings, co-founder and CEO of Filament, about his vision of a decentralized Internet. In last week’s episode, Jennings chatted with O’Reilly’s Mac Slocum a bit about a decentralized Internet in the context of the Internet of Things, and I ventured a bit deeper into the topic this week.
I asked Eric about the model — what would a decentralized Internet for the IoT look like and how would it work? He likened it to the Web:
We actually take a large portion of our model, our mental model, about a decentralized IoT from the early Web. If you imagine back in the early Web days — way back, mid-80s, early 90s — HTTP and websites had just started coming around, and they were originally focused and designed for academic research papers to link to each other.
Back then there was this entire, and there still is, there’s an entire open protocol stack that the Web runs on. Since any site could link directly to another site, it became very open and friendly, and there were all these wonderful things that emerged from — the Facebooks and Googles and WordPresses of the world were built on top of this standardized open reference platform.
What we like to think is, what would that look like if you took that concept and mapped it over onto the Internet of Things? What similar analogies to the Facebooks and Googles and WordPresses would we see if we had a truly decentralized and open IoT stack, and not necessarily one that’s full of silos and verticalized specific solutions to small industry segments.
Jennings described the current landscape as “a death by a thousand paper cuts.” He said the situation isn’t ideal and that it “feels very much like a feudalist mentality of getting the largest number of people on your side for your consortium to get people to use yours versus others.” I asked what the IoT would look like in 10 years if we continue down the path we’re on now:
If we go on the same trajectory, it will end up probably looking a lot like what the industrial Internet looks like today. You’ve got companies that make their own proprietary solutions and their proprietary protocols — even if they’re not proprietary, they’re not open or standardized. If you buy a product from one company, you have to use all of the products that go along with that company in order to make that solution work. You can’t bring in another product from another company and have them inter-operate very well. Even if they run the same standards, at the payload level or at the higher levels, they often don’t really communicate well together.
If we go down that road, I believe it’ll be a thousand of those types of companies trying to do their own thing, and they’ll probably be repeating a lot of the same solutions that others are doing. So, then you’ll get vying factions trying to solve the same problems rather than building some standards and openness where people can build apps on top of that or higher layer solutions on top of the more standardized underpinnings.
The biggest hurdles to a decentralized Internet, Jennings said, is a combination of for-profit companies realizing the market-share possibilities and trying to stake their claim, and the momentum cloud computing and big data are gathering in areas like wearables. The cloud worked for other things, why not the IoT? “The hard part,” Jennings said, “is to unlearn that a little bit. How do we unlearn that devices have to talk to the cloud? What if devices talked directly to each other? What if they never had to talk to the cloud? Maybe only once in a while they talk to the cloud? What would that look like?”
Jennings has a vision of a decentralized IoT stack to mitigate the problems a cloud-based, siloed, verticalized IoT would bring. He said the solution is to invent the future you want. That’s what he’s hoping to do with Filament Tap, an industrial machine-networking solution that grew out of the company’s first maker-focused product, the Pinoccio Scout. “Now we can take a 30-year-old manufacturing line and give it an API,” he noted. I asked how Tap fits into his decentralized IoT stack vision:
Alan Kay has a quote that I like because it’s so true: “The best way to predict the future is to invent it.” … It really does ring true to us, that if you see a future that you like more than the trajectory of the future that’s going right now in society or in life, then work your ass off to try to actually make that happen. Because if you can do that, even if you don’t make it all the way, you’ll get partially there. That’s really how things were invented up until now: someone thought there’s a better way and they built it. Maybe it went in a way they didn’t know or think, but it was probably better than the way they were going. The reason I mentioned that is because if we tried to build this grand vision of a decentralized IoT stack right now, it would fall on deaf ears because it’s not really solving a problem for people today, though we suspect it’s going to be a huge problem if we go down this road.
It’s like one of those things where, how do we build this cool decentralized future in a way that solves a problem for people now. The way we figured this is that the Tap solves real problems for industrial customers today — they want to be connected immediately so they can start knowing when their machines will break and get insights. … What if we use that as a Trojan horse to actually put a decentralized stack on the Tap, so that when we have these deployed at volume and a lot of companies using it … we just happened to slide in this decentralized IoT stack, and now this decentralized future is starting to form.
The decentralized IoT stack, Jennings said, is really about creating a reality where devices don’t require a central authority to operate. Industrial machine networking aside, what if your phone could connect to another phone without the need of a cellular tower? He noted the potential for device-to-device transactions and payments, but said that’s just the beginning:
It’s starting to feel to me a little bit like what happened in the Linux world, where there was this operating system that was ‘free,’ but more importantly it was open and it started to become very very popular for building applications on top of … It seems like this could be the Internet of Things equivalent to what the Linux kernel was to operating systems.
If a decentralized IoT stack is fully realized and done right, Jennings mused, we’ll see all sorts of things occurring that we can’t even imagine now: “I could see companies within an industry working together … the Caterpillars of the world and the John Deeres of the world letting their machines pay each other for data about loads a tractor is holding, or about how much capacity is left. … I have no idea what the future is going to look like, but it’s going to be pretty amazing if we can pull this off.”