Storj Labs CEO Ben Golub describes his latest startup, which tackles decentralized cloud storage, as “Airbnb, but for disk drives." And he has his sights set on empty disk space in desktops and data centers around the world. 

Storj, pronounced “storage,” this week launched a beta program and pricing for its Tardigrade Decentralized Cloud Storage Service, which will be available in January 2020. The company says it already has 14,000 developers on the waitlist. 

But in order for decentralized cloud storage to become mainstream and compete in the cloud storage industry, it needs to be easy to use and cost effective. To this end, Storaj's service is S3 compatible.

Amazon Web Services (AWS) is the most widely deployed public cloud, and “S3 compatibility alleviates many migration costs for users currently storing data on a centralized cloud provider and makes it extremely easy for developers to incorporate Storj as a storage layer, enabling migrations in minutes instead of days,” John Gleeson, VP of operations at Storj Labs wrote in a blog post. 

In addition to AWS, Storj runs in any of the other major public clouds including Microsoft Azure and Google, and only requires a couple lines of code to switch from storing on centralized platforms. 

The other part of Storj's plan to take on the centralized cloud storage market centers around Tardigrade's pricing. The company, in a blog post, claims its service costs less than half of the average starting price of the large, centralized cloud providers. Tardigrade also includes discounts based on bulk pricing, and it lets users add features like encryption and multi-zone redundancy at no extra cost.

Shawn Wilkinson, founder and chief strategy officer at Storj Labs, launched the company after struggling to find an easier way to store vast amounts of data from Twitter for a personal project he was working on. At the same time, he was also contributing to Bitcoin, which led him to connect the use of decentralization from the cryptocurrency ecosystem with cloud storage.

However, it can be difficult to locate a system of record — or where whole sets of data live — when it's distributed across multiple locations, which often fragments the data. To address this, Storj uses a deliberate erasure coding process that creates redundancy.

Erasure coding isn't a new technology. It's the same age as that scratched up Barry Manilow CD you can't bring yourself to get rid of. It's also what allows you to listen to that scratched up CD and hear Manilow's honey-smooth voice. Losing part of the disk doesn't mean the whole CD is ruined, and it's what Golub calls the "magic" behind the platform's durability and redundancy. 

On Storj's platform, files are split into 80 pieces and sprinkled into storage nodes. Of the 80 pieces, only 30 are needed to retrieve the files.

"It's a deliberate process that's managed," Golub explained in an interview with SDxCentral. "What sounds like fragmenting the data is actually a very tightly managed process that is audited and tracked so we always know where those ad pieces are."

This decentralized storage also makes the data more secure, the company claims.

At last year's Open Source Summit in Vancouver, B.C., Wilkinson explained how the Storj security model is like taking encrypted bits of sand and spreading them over an encrypted beach. “Instead of application data in a nice, manicured data center and hoping for no hack, you have to be more creative,” he said.

Wilkinson added that the failure of one computer has very little impact on the Storj platform. Because storage bins are spread all over the place with different users, power supplies, and different network addresses, Storj users get the functional equivalent of multi-regional storage for free.

“I do think that we're a reflection of what what people care about, and people are starting to care more about privacy, and more about security, more about value,” Golub added. “I think we're delivering that in a differentiated way.”