By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
It looks like data storage system vendors are in a bit of a bind these days. While the success of several high-tech housekeeping technologies won't put the brakes on the seemingly nonstop cycle of IT purchasing more and more storage hardware, it can sure slow it down considerably.
It could be the fallout from a couple of cruddy years economy-wise, or just natural evolution, but more storage managers are looking for ways to pare down the data they store to try to keep it close to manageable proportions. While most vendors responded with little more than lip service, 3PAR shrewdly saw an opening for its pioneering thin provisioning technology that provides a no-nonsense approach to using already installed capacity more efficiently. It wasn't just a cool feature; it was a savvy move by a startup vendor that ultimately forced most other vendors to include thin provisioning in their products.
With thin provisioning becoming well entrenched as a checkbox feature for most storage arrays, data reduction in primary storage (that's a mouthful, so let's just call it DRIPS) is taking center stage. DRIPS presents another approach-avoidance situation for storage vendors. I'm sure most would like to avoid it and just sell more disk, but if they don't roll out their own approaches to tweezing and squeezing primary storage data they'll run the risk of being left in the dust by their competitors. Credit NetApp and startups (upstarts?) Storwize and Ocarina Networks for getting the DRIPS ball rolling, and now everyone will have to get into the game.
While we're at it, let's add automated storage tiering to the list. Led (or pushed) by Compellent, which didn't invent the concept but certainly pumped some life into it, storage vendors are actually trying to help their customers put data in its place -- even if that place is on cheap disk.
So storage vendors might be just a little staggered by this one-two-three punch that lets storage managers buy storage later when it gets cheaper, buy less of it now or buy only inexpensive types of disk.
But storage vendors are a resourceful bunch and aren't likely to take this whole efficiency thing lying down. A number of them have already seized on another hot storage technology -- solid-state storage -- to counter the effects of possibly selling fewer storage systems and less-expensive disks to fill them. Until it gets much, much cheaper, solid state only has one role in the data center: handling the highest-performing apps in a company's portfolio. But array vendors are trying to tie solid state to automated tiering and other good housekeeping practices.
In my mind, I find it hard to link tiering with solid-state storage, but I've heard some pretty persuasive pitches that can make it seem only natural that you have solid state at the top of your tier and SATA below. And with automatic tiering moving your data around based on how IOPS-hungry it is at the time, it's logical to have the most expensive, highest-performing storage as the top tier. That all seems to make some sense on paper, unless the paper you're writing on is your company's check for a new array.
Any tiered storage arrangement has to have a top level of fast storage. But does it have to be that fast? Yes, it does, according to pitches I've heard recently from storage systems vendors. They were pretty convincing, and it all almost made sense until the price tag for that top-tier solid-state storage was revealed. The connection between automated tiering and solid-state storage seems contrived at best. It looks more like those vendors are conceding some ground to the efficiency techs while trying to carve out some new (and very lucrative) turf somewhere else.
There's no question that solid-state storage (even at its current astronomical price) has a place in many data centers, serving up data to applications that need the kind of performance that only solid state offers. And right now there's a tremendous amount of interest in solid state, but I believe that's mostly focused on the future of the technology when its price will be a lot less than 30 or 50 times that of spinning disk. Maybe one day solid state will also be an efficiency story, but right now it's just about pure speed.
BIO: Rich Castagna (firstname.lastname@example.org) is editorial director of the Storage Media Group.
* Click here for a sneak peek at what's coming up in the May 2010 issue.