Over at InformationWeek, Art Wittman recently revealed an important statistic from their recent 2011 State of Enterprise Storage Survey: 39% of the respondents indicated that 25% or more of their storage was direct-attached (DAS).
I’m actually not surprised by this. In fact, I expect this percentage is even higher among SMBs struggling to master server virtualization technology, which makes intense demands on storage capacity. Throwing more disks into a virtualization host looks like a good solution, especially when you’re concentrating on climbing the learning curve for server virtualization. It’s certainly inexpensive-in the short term.
In the long term, though, DAS is a dead end that can be very expensive. As Wittman points out, it’s “the hardest storage to manage, the least resistant to failure, and the toughest to fix when a failure does occur.” And even if you avoid disaster, at some point you’re going to run out of storage bays. Now you’ve got a server full of disks; where do you go from there?
Well, you can move your data to a file-level network-attached storage (NAS) box or a block-level storage area network (SAN) box. There are a lot of vendors that will sell you one, with as much physical redundancy as you think you need to keep your data safe (e.g., controllers, power supplies, fans, etc.). Eventually, of course, you’ll want two, plus storage management options like replication, mirroring, and the like, to avoid the single point of failure represented by a single box, and ensure that your data is really safe-which is a large part of the reason you moved away from DAS in the first place.
But that’s actually another dead end, although you may enjoy a longer run through the storage acquisition maze before hit the wall. The problem is, of course, that now you’re locked onto that vendor’s path through the maze, because the storage intelligence-scalability, data protection, high availability, management-is locked into the controllers in the vendor’s box. Same with your training investment. And all that storage you put in your host? Well, there’s always eBay.
No, it’s time to stop thinking in two dimensions, strap on a jet pack, and boost yourself out of the storage acquisition maze entirely with device-independent storage virtualization software. After all, your hypervisor has liberated you from worrying about whose server hardware you use: why not do the same with your storage hardware? By adding a third dimension to virtualization, device-independent storage virtualization software offers compelling business value (e.g., storage consolidation, increased utilization, reduced cost, improved availability) along with notable improvements in performance.
It’s really easy to get started. Storage virtualization software can run on an available server, run on the same platform as your hypervisor, or it can itself run in a virtual machine. You can also reuse the storage already installed in your virtualization server as part of an expanding pool of storage comprising whatever storage assets you have or add, including any storage from those tired old NAS or SAN boxes laying around. (Those assets probably include old servers, too: just add the software and disks and the result is a virtualized storage infrastructure that can adapt as needed to meet future needs.)
All the storage intelligence-tiering, mirroring, replication, and so forth-is there, ready when you need it, so you don’t need to “rip and replace” to upgrade your storage infrastructure. And, since the software will endure across multiple generations of storage hardware, you can get the most out of your hardware investments by moving older hardware into lower tiers as higher-performance hardware comes online, and preserve your training investments.
Best of all, you’ll find that the storage acquisition maze looks a lot simpler when you’re flying over it.
Photo by Jon Candy