This week, storage giant NetApp completed its buyout of SolidFire, one of the most successful early entrants in the flash storage market. In this installment of the IDG CEO Interview Series, SolidFire CEO Dave Wright talked with Chief Content Officer John Gallant about what drove the NetApp deal and what customers of the two companies can expect in the months ahead.
Wright, who will continue to lead SolidFire within NetApp, makes clear that flash is not some alternative storage technology any longer -- it's the primary storage option for cloud providers and enterprises as well. What differentiates SolidFire, according to Wright, is its ability to help customers build true scale-out infrastructure that solves the big problems IT is facing -- and that is only strengthened by the addition of NetApp's product portfolio. Wright also talked about customer successes and what lies ahead for the remaining members of the flash-only startup class of the past few years. (Spoiler alert: tough sledding).
Dave, let's start off with a little history of SolidFire. What did you guys try to tackle and how did you position yourself in the market when you launched in 2010?
What SolidFire saw six years ago when we started is that the way that enterprises and service providers are building their data centers is changing. The type of architecture that started with the public cloud is making its way into enterprise and service provider data centers. Virtualization has entered a new phase where the underlying infrastructure is being abstracted. And there is a new set of challenges facing data center administrators -- challenges related to scalability, automation, efficiency, multitenancy, performance management, all of these things that we didn't feel that traditional enterprise storage was really well-designed to address.
Quite frankly, a lot of the other early stage companies that were focused on flash and focused on performance weren't really focused on either. They were really only addressing one small piece of the problem, which was performance challenges, but there was a larger set of issues that we saw and that we ultimately wanted to address with SolidFire.
How is the enterprise buyer viewing flash now? What are you seeing in the market?
Flash is just becoming ubiquitous. It is no longer seen as the secret sauce you just sprinkle here and there. It's being seen as a mainstream storage media for anything that could be considered primary storage, and the days of people buying fast disk are kind of rapidly gone. As customers are replacing legacy systems, flash is the starting point. It's where they start the conversation. The assumption is that it's going to go on flash and really only the coldest, slowest data is going on disk if not moving entirely to the public cloud. Flash's role in the data center is broadening very rapidly, broadening from a performance solution to one that is just part of the mainstream storage world in the enterprise data center.
Is there anything that keeps people from committing to it more than they are at this point?
There's not a whole lot holding them back at this point. Price is certainly one that was a legitimate concern for quite a while. There was concern for it that was being pushed by the large vendors but really wasn't justified. Now, as the big vendors have gotten behind their flash strategies and have started toning down their FUD around price as a barrier, we've seen price really is not a factor and that the cost effectiveness of flash, particularly with data reduction technologies, is far superior to disk.
When you look at the barriers to adoption, customers are trying to figure out how to move past their legacy infrastructures and into next-generation storage systems. In many cases they face a broader set of decisions that they're making at the same time. It's not just about moving from disk to flash. Do I go into public cloud, do I go into hybrid cloud, do I continue with VMware, do I go OpenStack, do I look at hyperconvergence? There's a whole other set of infrastructure questions customers are wrestling with, and flash is being brought in as part of that discussion. But again, in most cases it's not the driving factor. It's a larger infrastructure transformation effort that they're underway with and flash is just part of the solution as they look at solving the bigger needs.
I want to go back into that differentiation question that I asked at the beginning. Talk about how you tackle problems differently than, say, Pure Storage?
Some of the early entrants in the all-flash space came at it from the perspective of how can we, as seamlessly as possible, replace disk-based arrays with all-flash arrays that have a similar capability set, a similar look, feel and function but are much faster? That was appropriate for a time in the market when customers were really just looking for faster replacements for their disk-based arrays. But the thing that really differentiated SolidFire is not only were we bringing that performance piece to the table, we were bringing a radically different architecture, one that was a true shared-nothing, scale-out architecture that gives a tremendous amount of flexibility and agility in building customer infrastructures -- flexibility they just don't have with a traditional dual controller, fibre-channel-connected array architecture.
[We have] the ability to start with a very small footprint, linearly scale capacity and performance, scale down the system, split a system into two, mix and match multiple generations of hardware and multiple generations of flash in a single architecture, seamlessly do generational upgrades on the hardware without any impact to the application, create a true unified pool of capacity and a separate pool of performance that can be provisioned through software across the data center and is deeply integrated into virtualization and cloud stacks like VMware and OpenStack.
Those are the things that really set us apart. It's not just another fibre-channel-connected, dual-controller array that just happens to be faster than the old one, it's a radically different architecture that brings new benefits along with the performance benefit of flash.
Quality of service (QoS) and multitenancy are two key features SolidFire touts. Can you talk about why those are so important to enterprise customers going forward?
If you look at the shift in the data center over the past 10 years and the transition going forward, it's very clear that the days of siloed data center infrastructure -- where you stood up a dedicated compute, networking and storage stack for each application -- are long gone. The pace of enterprise IT, the pace of application development is such that we need a more automated infrastructure, one where you've got pools of compute, networking and storage that can be dynamically provisioned and de-provisioned as application needs change and as workloads come up and go down and move from one location to another.
You just can't deliver that without multitenancy. 'Hey, I want to move this application to this data center but, by the way, I need to buy another storage system because it needs to go on its own platform'. You have to have true multitenancy, you have to have the scalability and the automation and security pieces to be able to deliver that.
But when you actually deliver that you also have to make sure that applications get the performance and consistency of performance that they need. That's where the quality of service piece comes in. You cannot consolidate hundreds of unrelated applications on a single storage platform, even a flash storage platform, without that kind of quality of service. You're going to run into ongoing noisy neighbor scenarios, performance headaches, inconsistencies that are going to drive the application owners mad.
Enterprises and service providers that want to enable that kind of true shared infrastructure, storage consolidation play have to have those multitenancy and quality of service features. Otherwise they're going to be back in the whack-a-mole world they used to live in, deploying a different storage array for every workload just to prevent them from running all over each other.