NetApp VP says storage virtualization overrated

NetApp's vice president of emerging products, Jay Kidd, discusses storage virtualization, competition with Isilon and NetApp's current identity crisis.

Network Appliance Inc.'s (NetApp) senior vice president and general manager of the Emerging Products Group, Jay Kidd, is on a whirlwind tour of North America. Last week he could be seen heading up a session at Storage Networking World (SNW) in San Diego on heterogeneous storage management; this week found him in our offices just outside Boston discussing topics, such as storage virtualization, clustered network attached storage (NAS) and NetApp's struggle with defining itself as it expands beyond its beginnings as a NAS company.

Virtualization has been a hot topic in the industry of late. What's your take as a NAS player on file virtualization?

Jay Kidd: In the file virtualization space, the NeoPaths, the Acopia's, Rainfinity -- none of them did all that well. If you put all those companies together, I doubt there's a dollar of profit in that entire set of businesses. We believe that there's a place for a virtualization engine at that layer, but mainly it's used for migration, typically lease-end migration. Some customers who are doing frequent migrations may buy the infrastructure, but it's not a big market.

More storage virtualization info

Users stand by NetApp -- for now


NetApp revenues fall $61M short of forecast


Goldman Sachs: Interest in NetApp SAN products wanes

SNW: Storage virtualization is no silver bullet, users say

Cisco's move to kill NeoPath shocks users, analysts

NetApp was partnered with NuView before Brocade bought them, and we didn't acquire that technology for a variety of reasons. Brocade did, and we're fine with that. We resell that actively -- in customer environments where they want to have a unified namespace across multiple filers, both NetApp and others. Then, having a host-based unification point makes a lot of sense. Within NetApp-only environments, our clustered system provides that unified namespace.

It's been suggested that the NetApp V-Series could be positioned to compete with the Hitachi Tagmastore in data center virtualization. What's your take on that?

Kidd: That is exactly what it is. We can put multiple different arrays behind it. Where we caution people is, if you move toward the level of having a single device, let's say it's virtualizing six arrays behind it, that device needs to have the performance and availability of the aggregate -- it has to be six times better than each array behind it.

So is NetApp taking a wait and see approach? Is virtualization going the way of ILM?

Kidd: The hottest thing going on right now in the data center is server virtualization. But what server virtualization is really about is consolidation. The same phenomenon existed in data center storage in the late 90s, and that's really what drove the movement toward networked storage. So now you've got consolidated servers, consolidated storage. Server virtualization has been a great thing for network storage, because VMware means the death of DAS.

Is it driving people to storage virtualization?

Kidd: Not necessarily. I don't think there's enough of a market for a single device that would support petabytes of storage behind it -- generally with storage, users want to have pools of sufficient scale that they have flexibility in terms of how they provision it, but they also want to mitigate risk by not having all their eggs in one basket. They want to have separate pools.

So how does that jibe with the utility view of storage?

Kidd: I think the essence of utility computing is that it's going to be laid out more like the Internet, where you have many-to-many infrastructure with many storage devices, many network devices, many host devices, rather than a big old bottleneck. There'll be intelligence in all of the end devices to know how to find their storage within that grid. This type of scheme is ultimately more scalable.

So where does NetApp fit into that vision?

Kidd: This is a new area of indexes and directories. There are several data management products on the market that create an index of your data -- we offer Kazeon's product. I think you'll see a lot of excitement around indexing -- you'll start to see different tiers of data classification. Indexing will be at the core of a lot of new products, and over time, it'll work its way into storage devices as well. This is where I think storage virtualization is focused too much on simplifying the physical. I think if you raise it up a level, you can simplify the logical, even with more complex physical pieces.

What kind of traction are you seeing with your GX clustered NAS product?

Kidd: We've had OnTap GX shipping for about a year, and there's good customer engagement with it. We've targeted high-performance compute environments, environments that have big data problems and need a lot of capacity and performance. It's not necessarily an enterprise focus now. We find generally database, Exchange and SAP users aren't looking for a clustered controller. They're looking for something a little more traditional, where the economics of the storage are compelling. but it's not a whole new different type of architecture.

Are you starting to compete with Isilon?

Kidd: I don't think we realized we were competing with Isilon until they said we were competing with them. We didn't really see them much in the markets and accounts we were selling into.

Doesn't that mean you were missing a whole customer segment then? Isilon has gone public on the strength of that market.

Kidd: That's a fair comment. I wish that we were big enough to invest in every single segment. About three years ago we really targeted the enterprise segment and wanted to build a business there, and we had presence in the technical segments. Did we leave some segments unattended? Yeah, we probably did. And Isilon has gone and found those segments and called our attention to them.

It's a challenge to balance, because even though we're growing 35% a year, we can't invest in everything we'd like to. We're doing more things right now -- we've broadened the focus in the past year or so to be back on the technical side again. One area where I think we haven't done much of anything is just building awareness -- a level of brand identity. Who is NetApp? Our technical buyers understood us, but as we've shifted into other markets, we're not as well understood.

So, who is NetApp?

Kidd: That's what we're figuring out now. We have ideas and will pursue ideas, and there will be disagreement and dissent even within the company about what we should pursue. We debate it very openly.

We know it'll be around simplicity. We know it'll be around data management in storage. We know it won't be limited to a single market. It's not a change in our perception -- it's an expansion. We're not just NAS -- that isn't going to be our tagline, obviously, but it's the truth. But the finer points of that identity are what we're going through an exercise now to really refine.

 

Read more on Integration software and middleware