The Shape Of Today's Storage Market

So often, I talk about industry directions and future scenarios.  That's to be expected, but my view is built on decent understanding of what the marketplace looks like today. 

IDC just released their storage market estimates for 2012 -- both Q4 and for the full year.  As I was going through the numbers, it struck me that many people might not have a high-level appreciation of how big the market is, what the major categories might be, and so on.

Although EMC continues to do very well as you might expect, this post isn't really about taking yet another victory lap -- it's more about drawing the broader picture of the storage industry as it stands today.

Consider it Storage Marketplace 101?

Before We Begin

All data here comes from the most recent IDC Worldwide Disk Storage Systems market estimate report.  I'll make this easy: in my opinion the IDC numbers have proven themselves to be the best numbers out there in terms of accuracy and useful segmentations.  EMC, for example, doesn't give IDC any data other than what's publicly available, but their numbers tend to sync very well with our internal ones over the years. 

Their models are very good.

The numbers reflect branded storage sold.  If you're an OEM, it's the name on the faceplate that gets counted, regardless of who built it.  With a little math, it's easy to back out OEM'd products and assign them back to the original manufacturer -- but it doesn't really change the big picture all that much when you do that.

IDC does a separate report on storage software; here we're counting storage boxen and any operating environment software required to make it work.

I've highlighted short excerpts from the latest IDC report, but -- if you have any interest at all -- you really should invest in buying the entire report and the data sets behind it. 

It's good stuff.

The First Category -- External Storage

Storage devices (disk and flash) can live either inside the server, or outside of it in some sort of external enclosure.  IDC calls the latter category "external storage", whether it has a network interface or not.  So this category would include dumb storage cabinets that supply only power and cooling -- as well as the more familiar networked devices .

External storage is a big category with decent growth.  IDC estimates that the entire market was worth $24.7 billion in 2012, and enjoyed a moderate 4.7% growth rate over 2011. 

Just for comparison purposes, IDC also counts three or more drives in a server enclosure as "internal storage" -- they don't count the server, just the drives as well as any required controllers.  It's not as big a number as you might think -- around $8 billion in 2012. 

Today, the vast majority of storage is sold externally, and not embedded in servers.

From my perspective, macroeconomic factors seem to be mostly responsible for growth rates, as you'd expect -- more storage stuff inevitably gets sold in years with robust GDP growth, less if there's a recession or economic uncertainty. 

But while 2012 might be charitably described as "uncertain" from that perspective, the reality is that you can only defer storage purchases for so long ...

More Interesting -- Networked Storage

Put some intelligence in your storage box, and connect it via some sort of network, and we have perhaps the most interesting IDC category: networked storage.

IDC estimates that in 2012, this market was worth $21.2 billion, and grew faster than the broader storage market at 5.4%.  This is not a new observation, it's been going on for many years -- networked storage in all its flavors continues to be where the action continues to be.

SAN vs. NAS vs. iSCSI etc.

The majority of the networked storage market is block-based: it presents a low-level block device abstraction to applications: FC SAN, FCoE, iSCSI, et. al. 

IDC describes this category as "Open SAN", and estimated it to be worth $15.4 billion in 2012, growing at a healthy 6.4% as compared to 5.4%. 

If you're interested, IDC tracks iSCSI separately.  It shows good growth, but off a much smaller base: $3.2 billion of iSCSI storage arrays were sold in 2012, with a decent 7% growth rate.

Not to get into a philosophical debate on the superiority of one approach over another, but you can see where people are voting with their money :)

NAS is an extended category through the IDC lens.  For starters, I believe that IDC counts any external storage device that presents a file (or object!) interface -- whether it's used for primary, secondary or archival storage.

With all of that, though, IDC's extended NAS category was responsible for around $5.7 billion and a modest 2.8% growth rate.  Block storage appears to not only be a much larger market than any form of NAS, but apparently growing faster as well. 

IDC has historically also broken down the numbers based on the OS the storage is plugged into: z/OS, Linux, Windows -- but no breakdown as of yet around hypervisors.  For example, external RAID storage associated with Microsoft Windows appears to be in rude health: $12.8 billion of storage was sold in 2012, with a respectable 11.4% growth rate.

Obligatory note: EMC is the clear market share leader in almost all IDC categories, with the notable exceptions of iSCSI (Dell is currently #1) and z/OS where it's IBM as you might expect.  In many categories, EMC is currently growing faster than the overall market and gaining share from competitors.

What Size Is Your Array?

While the IDC breakouts go into some good detail around specific vendor model numbers, geographic splits, etc. -- there are some interesting estimates around initial purchase prices for storage arrays.

IDC defines three broad price bands: one under $25k, one between $25k and $250K, and $250k+ with even more fine-grained detail when you purchase the service. 

When looking at the numbers, you have to consider that IDC is perhaps counting initial array hardware purchases, and not potentially counting subsequent expansions of either internal controllers or capacity.  In a world of modular and scale-out storage designs, this might not be as useful as it used to be, but it's still interesting.

In 2012, the entry category was responsible for ~$4.7 but has good growth -- over 10%.  The midrange category is larger, but growing more slowly: $11.4 billion with a humble 1.6% growth rate.  And the big iron is responsible for $7.6 billion and a very modest growth rate of under 1%.  EMC has a big lead in the top two price bands, and is currently behind Dell (and fighting it out with HP) for the #2 spot in the entry level category.

Before you jump to your own conclusions here, let me jump to mine: 

#1 -- I've been watching these numbers wax and wane for well over a decade, and I think macroeconomics are the primary force behind revenues showing up in one category or another.  Convincing someone to lay down big money for a big array can be challenging when there's uncertainty in the air.  Conversely, big arrays have historically been very popular when economic growth is strong.

#2 -- Small, entry arrays often beg to be consolidated into larger, more expensive ones as environments grow and mature.

#3 -- With newer scale-out architectures, the truth may be increasingly masked.  For example, in a very large Isilon cluster, it might have grown one or two bricks at a time.  Big iron built with smaller pieces ...

The Road Ahead

Anytime you look at one of these detailed reports from IDC, it's not hard to come up with new ways of potentially looking at the data. 

For example, what are the splits associated with various hypervisors -- the new "OS"?  Server flash is starting to become a meaningful marketplace -- how and where should that be counted? 

In a world of modular, scalable storage architectures, how does the notion of "price bands" carry forward? 

At some point, do we consider "cloud" a a distinct use case, separate from others?  If a service provider is providing a public storage service, does that make them a "storage vendor"?  Will there be a pronounced architectural shift to storage delivered as industry standard servers using scale-out software stacks? 

IDC has very well-informed views on many of these topics -- and more -- to be found in their most recent "Enterprise Storage Systems 2012-2016 Forecast Update" -- also worth buying in my opinion.

The longer you work in the storage business, the more interesting it gets :)

Comments

[Quantum]

Really I can only see good things coming from the market moving forward.

[Emily Drevets]

Great high-level overview of what's going on in the market, and it's certainly an exciting time to be watching the storage space.

Did you see this article from wired.com: http://www.wired.com/wiredenterprise/2011/11/emc-storage-king/2/

Interesting to see how much activity there's been in the past year. I wonder how these trends will affect market leaders down the road.

[Mila Kantcheva]

Chuck, great analysis here.

I just wanted to point out that your numbers don't quite add up: the $4.7B in entry, $11.4B in midrange and $7.6B in high end, do not make $24.7B.

This caught my attention because I've been puzzled by the low growth in entry storage we've been seeing lately (in other words entry storage is the number that's off)

[Chuck Hollis]

The numbers won't tie off until you purchase the entire IDC report :) Seriously, there's an extended taxonomy at work here, I'm just sharing the highlights.

[Mila Kantcheva]

My company does purchase the report and I double checked before posting. Again - not nitpicking at the numbers but wondering about the entry market dynamics - where the data paints a different picture.