CRN Exclusive: SolidFire's Wright On NetApp Integration, Cisco Partnership And The Future Of Flash

SolidFire: Not Just Another All-Flash Storage Company

NetApp's surprise acquisition of SolidFire early this year looked to most industry observers like a move by NetApp to catch up with the flash storage business. But as NetApp is already the world's second-largest provider of flash storage by market research firm IDC's reckoning, the acquisition appears to be really more about broadening NetApp's reach across a wide range of markets.

Dave Wright, SolidFire founder and CEO and now a NetApp vice president and general manager in charge of the SolidFire business, argues that SolidFire was built from the ground up to offer cloud-like capabilities for enterprises, service providers and hyperscale data centers. Indeed, SolidFire proved this point with the introduction this month of FlashForward Capacity Licensing, the industry's first program for selling all-flash storage as a perpetual software license with the underlying hardware as a separately invoiced item.

Wright recently sat down with CRN to talk about SolidFire's unique technology and how it and NetApp will impact each other's solution development.

What does SolidFire bring to NetApp? What's the difference in terms of technology and customers?

First is a new storage architecture, and one that's pretty radically different from anything else in NetApp's portfolio. NetApp's other flash products are fairly traditional dual-controller storage array architectures that fit very well in traditional data centers, but are not as well-suited to the types of next generation scale-up architectures that we see both service providers as well as a lot of enterprises trying to build going forward. SolidFire has a basic architecture that provides a great building block to go after that opportunity, which is one of the big growth areas in the market.

[The technology] we bring to the table around quality of service, around automation, around multitenancy for large cloud environments, is really intersecting with the sweet spot of where the opportunity for flash [is]. … The growth area going forward is going to be new-generation apps, cloud apps, cloud-like infrastructures, the place where SolidFire is very firmly focused.

Was that cloud-based focus built into part of your thinking when SolidFire was first developing its products? Or did that come about later?

It was really the whole idea from the beginning. A lot of folks think about SolidFire as a flash storage company because we do sell flash storage product. The real genesis of SolidFire was always around building a next-generation storage architecture for cloud-like infrastructures. I came from Rackspace, came from the experience of helping to build a large cloud environment and really seeing firsthand the challenges of doing that with traditional enterprise storage products that weren't designed for that type of architecture. SolidFire from the beginning has been focused on helping our customers modernize the data center infrastructure, move from the very siloed, hand-built model of the data center to a shared-nothing, pooled architecture [that is] software-managed, software-defined.

What is different in the SolidFire architecture that helps target the cloud-like infrastructure of the modern data center?

One is the simplicity of deployment in scale. … SolidFire has a much simpler model.

Think about a Lego brick model. You put in a couple of bricks, you need to grow, and you just snap in a few more bricks. It's four cables in the back and two clicks in the UI and you're done. In a rapidly growing cloud-type environment, that's the type of simplicity of deployment model you need, both for productivity reasons as well as to reduce the chances of error. You also look at things like the quality of service capabilities that we have. … In a traditional enterprise environment, the storage admins personally know every single workload on their storage array. They have a relationship with those storage workloads, and they know how to care and feed for them.

How does that differ in the cloud?

When you move into a cloud-like environment, you're delivering resources, you're delivering capacity, you're delivering performance. You're delivering it to a wide range of applications that you don't really know what they are, what they require, or what they're going to demand on a day-to-day basis. You can't care and feed for those things the same way you do traditional enterprise apps. You need a system where you can allocate resources out, you can have enforced SLAs by the storage system itself. And you can dynamically reprovision those resources very quickly, either because applications' requirements change or because one set of applications went away and another one came into that environment.

That's the type of capability that SolidFire provides that is really much harder to do on any traditional storage environment.

You're saying that traditional storage vendors cannot offer that type of capability?

No. We tried that firsthand at Rackspace. EMC and NetApp were both big vendors to us, and we looked at everybody else's products on the market. Again, if you look around the rest of the market, the vast majority of the flash architectures out there are still traditional dual-controller storage array architectures. They are updated versions of the architectures that have been dominant in the enterprise for the last 10 years. Scale-out architectures have made their mark in places like object storage. You certainly see scale-out object and file storage systems out there. SolidFire is really unique in bringing those types of simple, easy-to-manage, easy-to-deploy, scale-out capabilities to performance-oriented block storage.

Last year SolidFire introduced Element X , but you declined to call it software-defined storage. This year SolidFire is introducing FlashForward Capacity Licensing, which also sounds like software-defined storage. Can you apply that term now?

One of the reasons we've steered away from [the term is] because I think it means different things to different people. In our view, software-defined storage is less about how you're delivering the product and more about the management of the model. Is it a model where you are architecturally managing silos of hardware? Or is it a model where you have a pool of resources and in an intelligent software layer you're dynamically managing those resources? That's always the model that SolidFire has had. We've really always considered from that viewpoint SolidFire as software-defined storage.

How does that differ from other implementations of software-defined storage?

Other folks think about software-defined as a delivery model like, "Well, I buy my storage software from one person and I buy my hardware from another person and I put the two together." At the end of the day, we think that's a delivery model. There's a lot of traditional storage array architectures like ZFS that really don't have any special sauce in terms of cloud architecture. They're just a dual-control storage array that you can buy the software separate from the hardware. It's a delivery model, at the end of the day.

That said, we see the value in software delivery models for customers. There are, particularly in the hyperscale customer base, folks [who want to procure their own hardware]. We said, "Great. We're a software company at the end of the day anyway. We'll license you our software and you can manage your own hardware supply chain."

Did Element X succeed?

[Our goal] was never for it to be broadly adopted by every customer. We think most customers really aren't in a great position to manage their own hardware supply chain for a complex product like storage. We have had a couple of hyperscale customers adopt it. The goal was, No. 1, to address a set of customers who wouldn't buy from us any other way, and we've been able to do that. No. 2, learn about what customers are looking for in software-only or software-defined storage products. It created a lot of conversations with customers where they said, "Yeah, you know what, we don't like buying storage appliances anymore. There's a lot of negatives to buying fixed hardware/software appliances when it comes to storage. We like the flexibility and agility of the software-only model, but we're not quite ready to take on that supply chain."

That's what lead to the FlashForward Capacity Licensing program?

We [built] a program that has the best of both worlds. We can break apart the software and the hardware. We can sell you the software with a licensing model that looks and feels more like a software-only licensing model for that software, but we can still be the hardware supplier as well. We can take all the efficiency of our supply chain, all the volume purchasing that NetApp gets from our suppliers. We can pass that on to customers and just let them buy the hardware directly from us. So get the simplicity of appliance delivery with the flexibility of a software-only license model.

NetApp's traditional customers are more enterprise-focused, while SolidFire's have been more in the service provider and cloud-type environments. Does the acquisition bring SolidFire into more enterprise environments?

I think that may be a little bit of mischaracterization to both of our businesses. NetApp has always had a very, very strong presence in service providers. Relative to the size of their enterprise business, it was smaller just because the service provider market is a smaller market than enterprise. They've probably been … the top storage choice for service providers historically. They were the incumbent in a lot of the places that we were moving into in the service provider market. …

SolidFire started with a focus on service providers, but we broadened to the enterprise about two years ago, and we've had tremendous traction there. We do more than half our revenue in the enterprise now. It's certainly a very fast-growing market force even before the NetApp acquisition. With the [exposure to NetApp's] large enterprise customer base, we certainly expect to accelerate the adoption of SolidFire in the enterprise, but their service provider customers are also a ripe opportunity for us.

So SolidFire's technology fits well in the enterprise?

The technology itself is absolutely enterprise-class. In fact, what we've seen consistently as we go talk to enterprises is, we've really proven ourselves in the most demanding environments on the planet: service provider data centers, places where you have 24/7 revenue-generating infrastructures. [Being] proven in that environment really helps our credibility with enterprises. The feature set that we bring from the service provider world is absolutely the feature set enterprises are looking for. Enterprise IT is trying to become more like a service provider to the organization. That's what their business owners are asking them to do. "Hey, don't take my application and put it in a black box. I want you to become a service provider to me. Offer me a service catalog, offer me resources. Figure out how to do charge back and show back." They want something that they can really benchmark against the public cloud and they need [the] services to be able to do that.

As you look down the road, do you see any integration points between the NetApp and the SolidFire architectures?

We are not looking to merge the architectures. This is not something like [NetApp did when] they acquired a technology and merged it into Ontap. There are integration points between the different products and technologies in NetApp's portfolio. NetApp really has a portfolio strategy going forward. They want those products obviously to play nice with each other.

We've already done the first of those integrations with OnCommand Insight, which was their storage resource management reporting tool. ... But there's a number of other places that we expect to integrate as well, including things like the SnapMirror protocol to be able to seamlessly move data between SolidFire and other NetApp platforms. SnapCenter for unified snapshot management capabilities. We've integrated with their native Docker Volume plug-in so there's a unified Docker plug-in for SolidFire and NetApp platforms. There's a number of other places that we see integration opportunities as well [including FlexPod].

When will FlexPod configurations with SolidFire configuration be available?

We should see the first iteration of that in calendar Q3, which will be what's called a technology extension that allows FlexPod customers to add SolidFire to their existing FlexPod configurations. The full ground-up SolidFire-FlexPod designs will come out sometime after that. We haven't announced a specific timeline for those.

Will they be SolidFire only, or SolidFire combined with NetApp within the FlexPod?

It'll be just SolidFire for the storage, so no FAS. There'll continue to be FAS configurations of FlexPod, obviously, but there will also be SolidFire-only configurations targeting certain use cases.

Cisco had some issues with its own Whiptail, later called Invicta, all-flash storage technology, but now FlexPod will come with SolidFire for an all-flash version.

Yeah. We think it'll be a great solution. The intent at least with that and, again, this is something that we're still finalizing with Cisco, but the intent is to deliver that all on Cisco hardware. We have already qualified the CT20 platform as part of our Element X program for large customers.

Our intent is to bring it to market on [the CT20] platform. That'll be really exciting, because for the first time, FlexPod customers will have a 100 percent Cisco hardware technology that's still integrated with NetApp storage technology.

How will that differ from some of the other Cisco-based all-flash converged infrastructure offerings, like VersaStack with IBM or FlashStack with Pure Storage?

There's a fairly large difference between [them and what solutions vendors] bring to market jointly like VCE and FlexPod. They're simply an element of another vendor's converged infrastructure solutions. SolidFire had a converged infrastructure solution with Cisco before, Agile Infrastructure. But that wasn't a solution that [Cisco] was pushing.

FlexPod [is] pushed by both the Cisco and NetApp channel community. That makes it a very different go-to-market motion, compared to a lot of these other reference design-based [converged infrastructure] solutions. These are actual skued-up products that can be ordered directly from Cisco and NetApp channels. That, I think, makes it obviously a much broader and more compelling solution.

[The SolidFire converged infrastructure solution will also be] based on Cisco hardware, which none of the other solutions are today. ... That's going to make for a [much more] elegant scaling model around converged infrastructure than what you have today.

What are some things the market will hear about from SolidFire over the rest of the year?

One specific future announcement at least is Fluorine, the next release of Element OS, which will be in Q3. …

The big headline here is implementation of VMware VVOLs [Virtual Volumes]. VMware [environments account for] a little bit over half of our revenue base. We've seen a lot of interest in VVOLs, but not a lot of deployments. Part of that is that a lot of the early implementations of VVOLs by storage vendors were pretty weak. They were check-the-box-type implementations. They weren't fully integrated. They required you run a separate VMware appliance. They weren't highly available. They didn't really deliver the per-VM level of control that people were hoping or expecting with VVOLs. They just weren't mature, so people were sticking with their tried-and-true VMFS [Virtual Machine File System] implementation.

What else is new with Fluorine?

We're tripling the performance of Fiber Channel on our system. [And we've added] a technology called VRF Multi Tenant Networking. This is a virtual routing and forwarding capability that allows for complex, particularly service provider and telco, environments where you have multiple tenants that have multiple networks that may have overlapping IP ranges to allow them to still share the same storage system. This is pretty advanced. We were already, with our multi-tenant working capabilities, quite a bit ahead of the rest of the all-flash array market. This really puts us even further ahead for these really complex service provider and telco-type environments.

If NetApp hadn't acquired SolidFire, where would SolidFire be now?

We certainly were on track to an IPO and would have been looking to continue our growth and investment and have a public offering, likely in 2016. The reality is, the market is not particularly welcome to startup IPOs. There really hasn't been a true technology IPO in the first five months of the year so far. It's obviously tough times for a lot of startups who are looking at [an IPO] as their next step. ... I think SolidFire's backup plan always would have been to batten down the hatches and get to profitability and stay profitable until the IPO window opened up. I think we'll see a number of other startups look to take that path.

Could SolidFire have plowed through on its own without a large investment from companies like NetApp?

Absolutely. We had a path to growth, and we certainly had access to capital and other things we needed to continue to do that. It's certainly not fun to be in a market where the market is tightening and capital isn't as available as it used to be. That's always a difficult position, particularly for companies that have had easy access to capital. We certainly had a path forward, and it was a difficult decision, particularly at the time. When we looked at making this decision late last year, it wasn't clear what the IPO market was going to be this year. It looked like it might be a good market. Since then, it obviously has turned. It was a difficult decision at the time, but in retrospect I think it was absolutely the right one.

Do you miss not going through an IPO?

I don't think anybody loves the idea of going through an IPO these days, certainly not in the [Sarbanes Oxley] era and all the cost complexity going along with that.

When looking at those flash storage companies that had their IPOs, it looks like flash storage is not a magic technology for success. How has SolidFire done in terms of growth compared to the industry?

That's a great question. It really gets to the heart of why we've really never talked about ourselves as a flash storage company. We always saw the transition from disk to flash as simply that: a transition. It was inevitable. It was going to happen. Every vendor was going to participate in that transition. It's not that flash technology was exclusive to startups. Large companies certainly had flash offerings. NetApp's flash offerings in particular in the last 12 months have just exploded. Now they have over $700 million run rates. It was always clear that, while flash transitions would create some opportunities for startups, it would not be a forever and ever type differentiation. The big companies would, through acquisition or through building their own technology, in some way, shape or form, catch up.

How has that manifested in the flash storage market?

I think you're seeing some of the effect of that in some of the flash companies that have gone public, like Violin and Pure, where they're challenged. Their lead and their differentiation of simply being a flash technology company is gone. There's a lot of flash technology in the market.

For SolidFire, we think that speaks to the heart of why it's important that we never differentiate ourselves based on being flash. … Everybody's got flash. That's not what's important over the scale-out, the quality of service, the automation, all those pieces. Those have continued to remain differentiated. We have not seen other competitors implement similar capabilities. That differentiation remains, and that's led to very consistent year-over-year growth for SolidFire for the past several years. Obviously we expect acceleration of growth now with NetApp and the go-to-market resources at our disposal.

Prior to the acquisition, SolidFire was 100 percent channel, right?

Not quite 100 percent. We started our channel program about two years ago when we launched our enterprise business. About three-quarters of our business goes to the channel. Pretty similar percentage to NetApp overall.

Did many SolidFire channel partners deal with NetApp in the past?

They did. I think somewhere around two-thirds, and maybe even higher than that, of our channel partners were NetApp channel partners as well. For them, a lot of overlap, easy transition. I think [it's] really indicative of the fact that NetApp's channel is interested in SolidFire. A lot of those channel partners had already come into the SolidFire camp saying, "Hey, we need something complementary to our NetApp technologies. SolidFire is a good fit there."

Some of them were quiet about it because they didn't want to offend NetApp.

I understand, but they kind of speak with their feet.

What does SolidFire bring to NetApp's wider partner base?

First is a new story to tell. I think NetApp has been beating the Clustered Ontap drum for a while now. I think they continue to build and iterate and have a great product there. The NetApp channel partners have been looking for, what's the next story? What's the next opportunity? What's the new technology that we're going to bring to customers to get them excited about NetApp again? SolidFire certainly gives them that opportunity, where they don't have to [use a startup's product to fill a gap in NetApp's line]. Now they can say, "Hey, NetApp's bringing something to the table that actually is very new and exciting, offers new capabilities and new architecture to consider. Let's see if this fits for you." It enables them to have a conversation without having to go outside of the NetApp comfort zone.