Qumulo storage IPO possible, CEO says

Six years after entering the scale-out file system market, Qumulo is no longer a startup, but it isn't a major vendor either. The next year or so could make or break Qumulo as it navigates without founder and original CEO Peter Godman -- who left the company in late 2018 -- and prepares to possibly become a public company.

We recently spoke to Qumulo CEO Bill Richter about the company's plans, post-Godman and pre-initial public offering (IPO). From a technology standpoint, Richter said, the future of Qumulo storage is tied to the cloud. Enterprises desire file storage that spans heterogeneous local and cloud environments, although object storage has been dominant in the cloud for unstructured data, he said.

"File as a data construct is incredibly important," Richter said. "We believe that because of the thousands of conversations we've had with customers and the variety of things they do with file data."

Qumulo claims hundreds of customers use its Qumulo Core scalable file system. Qumulo Core is sold on branded QC-Series hybrid, P-Series all-flash NVMe and K-Series active archive arrays, and through OEM partnerships with Dell EMC and Hewlett Packard Enterprise. Google Cloud Platform (GCP) recently certified Qumulo storage, adding to Qumulo's earlier support of AWS. The Qumulo File Fabric file system runs inside GCP and can replicate local file data between Amazon and GCP.

Bill Richter

Richter, who became Qumulo CEO in 2016, is no stranger to the scale-out file market. He was an executive with Isilon, a NAS startup acquired by EMC for $2.25 billion in 2010 and rose to president of the EMC Isilion storage division. Former Isilon engineers Godman, Aaron Passey and Neal Fachan developed Qumulo storage and launched the company in 2012. Fachan, Qumulo's chief engineer, is the last remaining founder with the company. Passey left in 2017 and is now principal engineer at Dropbox. Godman resigned last year, two years after turning over the CEO reins to Richter and taking on the CTO role.

Richter gave us his take on the management changes, what's fueling demand for Qumulo NAS in the cloud, a possible IPO, and how he intends to beat clustered NAS giants, such as Dell EMC and NetApp, as well as well-funded data management startups.

Qumulo closed a $93 million funding round in 2018 and has accumulated more than $220 million in investment since its inception. What are you planning to do with that cash? How soon will you be profitable?

We don't blow money. Snoop Dogg will not be at our kickoff event.

Bill Richter: We don't give out forecasts on cash flow breakeven. We very much think Qumulo [will be] a long-term independent company, and the public capital markets are a possible path for us. An IPO is a couple years out, but I would love to take Qumulo public one day.

For sure, we're in investment mode and not in the black. Unlike many companies in this sector, Qumulo is taking a long-term view. We don't blow money. Snoop Dogg will not be at our kickoff event. We're not only on a mission to help customers with their file-based data, we're on a mission to build a sustainable business. We have a lot of cash on our balance sheet and not one dollar of debt.

Why did Qumulo founder and your predecessor as CEO, Peter Godman, leave his CTO job at the company? And what impact will that have on the Qumulo storage roadmap?

Richter: Pete is a dear friend, and I speak with him probably once a week. For him, it had been a seven-year journey. He transitioned first out of the CEO role to the CTO role, and then was in that role for a couple years. Then it was time for him to move on for a lot of personal reasons. Pete's leaving certainly will not impact the vision for where we're taking the company. There's no correlation there at all.

What has been the biggest change in scale-out file storage during the six years that Qumulo has been in the market?

Richter: Our core belief system is that file storage in the public cloud is really important. In the modern world, that means we need to be software-defined, so you can take our file system and run it reliably on third-party hardware platforms. It means we can adopt newer hardware form factors much faster.

The moment NVMe became commercially available, and economical, we adopted it as our all-flash media. The fact that we don't have an NVMe RAM card as the heart of our architecture allows us to run native file services very quickly in the public cloud.

You recently launched Qumulo CloudStudio and other file services in the Google cloud. Why did you select GCP as your second supported public cloud?

Richter: We have a lot of customers in media and entertainment, so we launched Qumulo CloudStudio in Google to run the entire end-to-end suite in the public cloud. We picked AWS first in 2017 because of market share and other obvious reasons. Then we spent a lot of time polling customers [about] which cloud to support next. Google is doing some really interesting things that are vertical-specific to our customers.

Given Qumulo's emphasis on the cloud, won't you need to enhance support for object storage to complement your file storage?

Richter: Think about it like coffee and tea: They're both important, but it's a matter of preference. We have the capability to support object workflows now. We're not object-opposed at all, but the bigger gap in the market -- and what customers are begging for -- is modern enterprise-class file storage.

People sometimes conflate object storage and cloud. We don't believe that's true. There are customers that want to use file in the public cloud, but the only options they've been given were very old, limited and certainly not cloud-native or software-defined.

What are the biggest obstacles facing Qumulo storage in the cloud?

Richter: What we're working on the most is scale. We're effectively hiring go-to-market people as fast as we can. Our limiter there is [available] talent. We have a very specific culture at Qumulo, so as we scale the business, we have to make sure we hire people that believe in our mission.

And boy, the market is really noisy right now. Legacy vendors talk about the cloud, but what they mean is a box that supports Amazon S3. We don't consider that to be the cloud. You hear some backup vendors suggesting they should be primary storage vendors, and that's a terrible idea. You've got all-flash block companies that support NFS and call themselves an enterprise file vendor, but we know that won't work. It's our job to cut through the noise ... to clarify what we believe and why we believe it, and let our work speak for itself.