That boring stuff -- storage -- is where we keep all of the interesting information we care to own.
The size and growth of the storage market can often be a rough proxy for economic macro-trends such as aggregate capital investment, shifts to digital business models, and more. More reasons to keep information around == more demand for storage technologies.
A recent IDC report claims that enterprise storage revenue growth will be notably more muted than in past years -- now forecast to track to a respectable (but modest) revenue growth rate of 4.1% through 2017. The report points at the adoption of familiar storage efficiency technologies -- compression, dedupe, thin provisioning, etc. -- as the primary culprit of reduced expenditures by IT organizations around the globe.
While there's no denying that many of these technologies are becoming widely adopted, I think there's more to the story. And I'm guessing that when 2017 rolls around, the picture may turn out differently than IDC is forecasting.
Why Does This Matter?
As a positive case, consider the rapid innovation in smartphones. For those of you familiar with the downward spiral of the tape market, you'll recognize that not only is there a paucity of innovation, but competitive alternatives are drying up very fast indeed.
If you're a heavy user of any enterprise technology, you benefit by the market being in rude health ...
There's More To The Story
We are all fortunate that the last few years have brought us all sorts of ways to buy less storage capacity. Thin provisioning means less waste. The advent of space-saving copies added more benefits. Throw in compression and deduplication for even more capacity savings. And, finally, flash storage means we can now scale performance without scaling the number of disk spindles.
All good. And, taken as a single thought, one might assume that's the entirety of the story: there's now a critical mass of space-saving storage technology out there, so the aggregate demand curve is shifting downward somewhat.
But, if you hang around the storage world, you might ask yourself -- why now? After all, this list of technologies have been widely available for some time, in some cases for many years. Why are we seeing a demand shift now?
IDC discusses this in a back-handed way, but I think the thought deserves more attention: operational models are changing as well. Service catalogs are becoming the norm, automated policy-based provisioning is more prevalent, detailed reporting back to the business on who's using what -- IT organizations are starting to look more like service providers (or cloud, if you prefer) than the legacy have-a-hunch-provision-a-bunch approach.
Indeed, I've seen customers go through the whole journey, many times.
It starts in a familiar place -- e.g. the customer is concerned that they're spending too much on storage. We, as vendors, respond with different technologies and alternatives to help them rectify that problem.
Many presentations and proposals ensue.
But as they start to consider what it might take to actually implement said alternatives, there's the beginning of a realization that -- just perhaps -- their processes and workflows aren't the most efficient. So they start down the road of refashioning themselves to provide attractive storage services vs. simply standing up and running the arrays.
Some may claim "well, it's the cloud!". There's an element of truth in that. There is now a wide array of external enterprise-class storage services that can be easily compared against internal alternatives. And that can start a very interesting discussion indeed -- how do we compare internally?
Hint: we've found that a decently-run IT shop can usually deliver storage services for significantly less than any external provider.
And that phenomenon -- timing-wise -- is visibly starting to kick in. All good.
So that's the supply side: more prevalent use of efficiency technologies, coupled with a proliferation of re-tooled IT functions -- or a move to an external service provider if that's in order. But what about the demand side?
Maybe I'm A Storage Optimist
Sorry, but from where I sit I see a veritable tidal wave of new storage demand that may not be fully factored into the IDC forecast.
Call it big data, call it big content, call it transactional storage -- the wave has many components, all working together to keep demand well ahead of supply through 2017.
One example: the first time I saw a production big data cluster that was larger than the entire enterprise IT component, I was surprised. I'm not anymore. Once any organization gets seriously addicted to analytics, it drives a substantial incremental IT spend on supporting infrastructure, with storage being the primary beneficiary.
Another example is the fast-rising popularity of enterprise sync-and-share. Based on what I've seen with EMC's Syncplicity, it is very addictive stuff indeed -- you want to put *everything* in it. Most enterprises are choosing to host the content behind their firewall, driving some very big capacities indeed. Ten thousand enterprise users at 50GB each is 500 usable terabytes -- and that's usually in addition to what you might have already spent on VDI or file sharing.
Looking just a bit farther out, we've got the Internet of Things, the Industrial Internet or however you'd like to refer to it: data generated by sensors vs. people. And the wave just got an order-of-magnitude more interesting.
For those of you in larger enterprise IT settings, how much potentially valuable log data does your environment generate every day?
The Historical Angle
I've worked at EMC for almost 20 years now. The disk drive of the day in 1994 was a 3GB Seagate.
Every time a bigger drive came along, there tended to be some fretting as now per-unit capacity costs had essentially dropped. You can look at the advent of space-saving technologies + improved operational models as an analogous effect: paying effectively less for what you use.
3TB drives (1000x) are now the norm, and bigger ones are certainly coming. Every time a bigger drive came along, people found a way to fill those up, and much more. As a result, aggregate storage expenditures go up every year, unless there's a nasty recession going on.
Put differently, the historical evidence shows that storage capacity is positively elastic: as costs drop, more is consumed and expenditures increase.
The Shift To Software
Let us not forget that all of the aforementioned space-saving features are the result of clever algorithms, more effective management and better integration, e.g. software. Smart software (properly used) leads to consuming less physical capacity -- and that's been true for a while.
As we move to software-defined storage, the opportunities become even greater for dynamic efficiency: not only capacity, but performance as well as protection. And that's *before* we examine the potential of software-based storage stacks running on commodity hardware.
There's No Arguing With The Numbers
IDC's analysis -- derived from the results of all the storage vendors -- is hard to argue with. But what we can debate is the shape of the curve in the future.
Predictive models are tricky things indeed, especially when there are structural shifts in play.
I suppose I'll have to wait until 2017 to figure out whether I was right ...
Like this post? Why not subscribe via email?