NetApp marked its 25th year of business in 2017, highlighted by its launch into hyper-converged storage based on...
SolidFire all-flash arrays. The NetApp market strategy is in transition. Layoffs were part of recent cost-cutting measures, as well as an earnest push to sell more all-flash storage.
Founders Dave Hitz, James Lau and Michael Malcolm started NetApp in 1992. The company started out selling Unix workgroup-level filers to programming teams designing software using NFS, and then it became the leading NAS supplier.
Twenty-five years after helping to start NetApp, Hitz, the executive vice president, is the lone remaining founder still at the company. We spoke with him about the NetApp market changes over the decades as NetApp has adapted to technologies such as cloud computing, flash storage and hyper-converged infrastructure (HCI).
There are changes going on in enterprise storage today, including continual data growth and heightened customer expectations for fast storage. How do these changes compare to what NetApp experienced in its early days?
Dave Hitz: NetApp has been through radical transformation multiple times. When the internet came along, we basically ripped up our roadmap. When VMware hit, many people thought that meant everything would be virtualized, not just servers. That view was only half right.
We thought server virtualization was awesome, but we also knew that the data challenges were immense. We dug in to start helping customers with managing those data challenges. Much of our growth was fueled by how well we partnered with VMware. Even after EMC bought VMware -- which scared us, I'll admit -- a lot of analysts said that NetApp had a better culture of partnering than EMC did.
The previous big transformation was being a tech and internet company. Early on, our foundational growth was in the internet space in the dot-com boom. When that collapsed, we were in a world of hurt. In the first six months after the crash, not one of our 10 biggest customers bought a storage system from us. Our stock went from around $150 down to about $5 or $6, like a plane in a nosedive.
Fortunately, before the downturn, our CEO and CFO were paranoid and said, 'We've got to be more diversified.' We picked five vertical areas, including banking and financial services, major manufacturing, medical and government. We hit $1 billion in revenue and then dropped to $800 million after the dot-com bubble. After three years, we were back to $1 billion in revenue.
The first time we hit $1 billion, the NetApp market was 70% tech and internet and 30% of the new style of enterprise storage. The second time we hit $1 billion, it was flipped the other way. That was an enormous transition for us technically, as we started to add SAN failover and super-high-reliability stuff.
Despite the rapid growth in data, the enterprise storage market is not growing much these days. How do you avoid another nosedive?
Hitz: If you look at the overall on-premises storage market as a whole, growth is flattish. But within that trend are some radical changes. One is the flash style of storage. We started late in all-flash, but our growth been amazing. Of the top five storage vendors, NetApp has the fastest growth in flash, according to the analyst firms.
Our share in the flash space is higher than our share of the overall enterprise market. If we can hold that position, we should be able to emerge out of the transition to flash with a higher share than when we first started, even if the overall storage market remains flat.
Another big opportunity is in what I describe as the next-generation data center: lots and lots of commodity white box 1U systems, with Ethernet cables running to a switch at the top of the rack. Increasingly, this is the zone of hyper-convergence, which is why we announced our HCI product this year based on SolidFire all-flash arrays.
As with all-flash, NetApp was late to the hyper-convergence market. How does SolidFire stack up to established disk-based HCI platforms?
Hitz: My mental image of SolidFire is that the data just flows between the nodes like water. Picture a four-node minimum of SolidFire nodes. If you attach another node, SolidFire does a little math to let the data drift from a full node into a node that isn't full. SolidFire software-defined storage automatically rebalances the system when you add a node.
Nutanix and other HCI vendors focus so much on virtualization side because VMware costs so much. Many customers would love to replace VMware. But they aren't solving a new problem. For us, we can partner with VMware to solve the back-end storage issues within the context of hyper-convergence.
A big part of the NetApp market strategy revolves around the cloud. What have you done to try and separate your cloud services from other enterprise array vendors?
Hitz: In 2014, we launched OnTap Cloud to run in Amazon Web Services. It's a version of OnTap running in the cloud that uses Amazon compute and storage. There is no physical system or cost of goods. All the revenue is from software. Customers use Cloud OnTap to migrate their data seamlessly back and forth. We also provide the enterprise storage for Microsoft Azure.
That is exactly what we are trying to [do] now with NetApp Data Fabric. What we're doing goes way beyond selling a storage array. Even though the bulk of our revenue is from storage systems, don't think of [us] as a storage system vendor. The enterprise on-prem business isn't going away, but at the same time, the center of gravity is moving to the cloud. When they buy on-prem storage, customers are asking lots of strategic questions about the cloud. One of them is: 'Which vendor can help me move to the cloud when I’m ready in a year or two?’ We want NetApp to be the obvious answer.