It’s about the water, not the bucket. The money, not the safe. And for Qumulo, a Seattle startup that’s just raised $40 million, it’s the data, not the storage.
Qumulo—founded in 2012 by early employees of enterprise storage provider Isilon Systems, which was acquired just over four years ago by EMC for $2.5 billion—is part of a new generation of enterprise IT infrastructure companies growing up in Seattle. They’ve attracted lots of capital in the last three years.
Qumulo’s Series B funding round was led by Kleiner Perkins Caufield & Byers (KPCB). Previous investors in the company’s $24.5 million Series A round, including Highland Capital Partners, Madrona Venture Group, and Valhalla Partners, also joined this round, as did Sujal Patel, Isilon’s founder.
Patel, KPCB partner Wen Hsieh, and Madrona’s Matt McIlwain are joining Qumulo’s board of directors.
So what is Qumulo up to with all that capital, besides paying its staff of 90 and counting? The company has kept a tight lid on its specific plans. Co-founder and CEO Peter Godman shared more in an interview last week, but was still coy, saving the details for a broad product introduction planned later this year.
Let’s start with the context. The rapid growth of data in the 21st century—created by individuals, devices, sensors, business transactions, large science experiments, and seemingly everything else—is well known. “We live in a virtual world. The stuff of the virtual world is data. And we never, ever throw data away,” Godman says.
Enterprise storage in the last decade—when Isilon came up—was about helping companies manage hundreds of terabytes and then petabytes of storage. Vendors provided large businesses bigger and bigger buckets—often in the form of proprietary storage servers and other expensive, custom hardware that quickly filled to capacity—but they haven’t been very good at helping people understand what’s inside the buckets, Godman says.
“The next decade we believe is really about helping people store billions and trillions of files, and manage them, and understand exactly what they have,” Godman says. “Managing lots of storage and managing lots of data—they’re really different problems.”
Data management, he says, means being able to answer questions about how and why an organization’s data is growing, where it is, who’s using what, what can be put in the cloud, what can be backed up?
“All of these questions are very complicated to answer, and people spend a lot more time answering them now,” Godman says.
A large organization wouldn’t tolerate spending tens or hundreds of millions of dollars on something like gasoline without a clear statement of accounts describing where it’s located and how it’s used. “Yet, we’ve found that that’s largely the status quo” in data storage, Godman says. “Either people invest a great deal of their own labor into starting to understand their data footprint, and how it’s getting used, or they don’t know.”
The company’s software-defined, network-attached data storage system takes advantage of commodity hardware and the availability of NAND flash memory—the stuff in USB drives and mobile device memory cards—at large scale and lower cost, he says. While flash memory has enabled high-performance storage applications over the last five to 10 years, Godman says there’s an opportunity now to use it to make data management easier.
And, as is expected today, the software is ready to run in any computing environment, and can scale-up to handle the largest data storage needs.
“Now our architecture is based entirely on commodity hardware,” Godman says. “That means our software will be able to run on appliances, inside virtual machine instances, inside cloud instances, et cetera.”
Godman says Qumulo conducted upwards of a thousand
