Stay on Top of Enterprise Technology Trends
Get updates impacting your industry from our GigaOm Research Community
There are several efforts to build a more decentralized internet — from attempts to build peer-to-peer wireless networks like Commotion or Open Garden to similar efforts with P2P browsers, such as BitTorrent’s Project Maelstrom. And thanks to blockchain, the decentralized data transaction processing technology behind Bitcoin, we’re now hearing more about how we can build a decentralized network for the internet of things.
This isn’t just some geeky effort to test technology or to fight “the man.” There are legitimate business and practical reasons that one might want to operate a peer-to-peer IoT network, as Eric Jennings, the CEO of Filament explains on this week’s podcast. Jennings’ company is building a module for industrial customers that includes a sensor as well as connectivity that will connect the sensor to a cloud if customers want it, but more interestingly to a mesh network of other sensors.
Filament is developing a group of technologies that will allow companies to deploy these sensors in the field without needing some kind of cellular or wireline backhaul to the internet. This is great for remote locations, where such connectivity might be hard to come by but machines still need to communicate with each other, or for people toting tablets. Another reason decentralization works is because it cuts down on costs of sending and storing data in the cloud.
While those costs are low, it may not make sense to store every bit sent over the 15-year life of a washing machine in the cloud, especially if manufacturers can’t come up with a business model that keeps the lights on in that data center. But the most interesting element of my conversation with Jennings came after we discussed the pros and cons of the first three elements of his technology group — blockchain, Telehash and BitTorrent — and dove into the tricky business of managing a device’s identity.
Jennings said the question his team was trying to answer was, “How could we have DNS for these devices without servers?” The answer was blockname, a proposal that uses the existing domain name system, but falls back to the decentralized system when it fails to find the device it’s looking for on the public DNS.
In the blockname trial that Filament has set up, when this happens, the request is routed via Telehash to the blockchain, which then looks throughout the chain for the requested IP address (or identifying number) of the device. I asked Jennings who handles that initial registration of the device in the blockchain, because in a decentralized environment, getting the correct and authenticated version of a device’s identity could be a challenge.
Jennings isn’t sure yet, but he hopes that the consensus elements of the blockchain technology help — basically if people disagree with you, they need more computing power to change the status quo. So to get the details on blockname, which start late at about 46 minutes in (Jennings’ section starts at 22:00), or just to hear Kevin Tofel and me riff on privacy and Samsung’s smart TVs, listen to the show and stretch your brain. When you’re done, feel free to send us a question before the end of the month and you’ll be entered to win a Chamberlain MyQ connected garage door opener.