Why the economics of cloud storage aren’t always what they seem

Gleb Budman has built his cloud-backup company, BackBlaze, into a profitable venture that houses petabytes of data in total and costs users only $5 a month for unlimited capacity. The company’s open source designs for building massive-capacity, dirt-cheap “pods” housing 180 terabytes apiece are so popular they even attracted interest from the CIA. But when Backblaze started, Budman told us on this week’s Structure Show podcast, people thought he was crazy.

Here are my favorite quotes from a very quotable episode, in which Budman shares his thoughts and insights on everything from the economics of storage to his presentation at the annual FailCon event. You’ll certainly want to listen to the whole show to hear all that Budman has to say, though.

[soundcloud url=”https://api.soundcloud.com/tracks/116861363″ width=”100%” height=”166″ iframe=”true” /]

Download this episode

Subscribe in iTunes

The Structure Show RSS Feed

On what inspired $5-a-month unlimited storage and, thus, the pods

“There was this huge bell curve around 5 bucks, where people said for 5 bucks it would just kind of be a no-brainer,” Budman said. “We said, ‘OK, great. We need to offer this unlimited online backup service for 5 bucks a month. How do we do that?'”

The answer, after determining that would cost $5 a month to store 30 gigabytes of user data on Amazon Web Services, and thousands of dollars per terabyte for enterprise storage arrays, was figuring out how to take advantage of $100 1-terabyte consumer-class hard drives.

“We started screwing around with hard drives and trying to figure out how we could get a bunch of hard drives connected to the internet for the lowest cost possible,” Budman said.

Gleb Budman Backblaze Online Backup Adrian Cockcroft Netflix Structure 2013

Adrian Cockfroft, Netflix (left), Gleb Budman, Backblaze (center) and Derrick Harris, Gigaom (right) at Structure 2013. (c) Pinar Ozger

On consumer skepticism once $5 a month became reality

“When we offered this unlimited service, we actually had potential customers that were nervous because they said, ‘Well, there’s no way they can afford to do this, which means either they’re just burning through VC cash and they’re gonna go out of business when the VCs stop giving them money, or they have some scheme. They’re gonna take all our data and sell it to someone,'” Budman said. “And we’re like, ‘No, we’re actually profitable, this works.'”

That skepticism led to the first of Budman’s blog posts explaining how the storage pods were built and how much they cost. After that, he said, people began to believe the company really was for real.

The NSA wants to store a yottabyte of data, really?

Budman read the stories several years ago claiming the NSA wanted to store a yottabyte of data, and he felt compelled to do the math explaining why that is implausible. His findings on the cost and footprint of storing so much data, even with the Backblaze pod design: “It was way more than the entire U.S. GDP. And it would take all of Manhattan, and I believe Delaware as well, to fill up with data centers full of storage pods to store that much data.”

Maybe that post made the rounds through the intelligence agency, because about two years afterward, Budman said, he got a visit from the CIA asking Backblaze to build it a massive system for its new focus on data collection and mining.

An aisle full of pods in the Backblaze data center.

An aisle full of pods in the Backblaze data center.

Admitting failure is a good thing

Budman has been involved with or at least attended the FailCon conference for years, he said, because it’s such a great way to learn from people’s mistakes. This year, he presented: “The talk that I shared was that a million-dollar marketing budget is not necessarily going to make it all OK.”

“A year ago, we decided to raise a $5 million round of funding for a variety of reasons, but one of them was we wanted to see if we actually spent money on marketing, which we hadn’t until that point, would the company grow faster,” Budman explained. “And over this last year of doing that, what we found was, um, it’s hard.”