Storage Automation: Oxymoron?

A new buzzword has popped into the storage lexicon, straight from the marketing department: automation.

It's a term not often associated with storage networking. Indeed, it seems many storage pros would go to great lengths to avoid automation, smacking as it does of loss of control and the fearsome potentialities that conveys.

But automation within IT in general is on the rise. BladeLogic's IPO and the proposed sale of Opsware to HP for $1.6 billion demonstrates its growing popularity. And several SRM vendors, most recently Onaro, are asking, "Why can't storage get in on the act?"

There may be several reasons, actually. As Yogesh Agrawal, Symantec's senior director for storage management, told Byte and Switch last week, storage is pigeonholed in large enterprises, which are typically the only ones who can afford automation. "If you look at a typical operational team, they are in silos," he says. And in silos they'd like to stay. Storage managers just aren't, as a species, comfortable with someone else's driving. Only from a well-protected silo can they maximize control and ensure that no one from outside messes up their highly specialized tasks.

A related reason why storage may not get automated anytime soon: The folks who buy IT automation aren't the same ones that purchase storage. Vick Vaishnavi, director of product marketing at BladeLogic, is blunt: "About the only time a customer wants to talk about storage is relative to servers and applications." Sure, his company has partnered with Onaro and others, but the main focus for BladeLogic remains on servers, not storage. The same may be said of Opsware and a range of other automation suppliers.This isn't to say that some storage management functions won't be automated. Tasks like provisioning and capacity planning are still labor-intensive and largely manual. (Think Excel sheets and plenty of feet on the street.) Reducing even a bit of their bulk could save oodles of time and money.

As storage automation and management improves, it's likely that a select group of products will join a number of IT management wares in a consolidated data center NOC. The integration level will be vendor-determined; history has proven it's unrealistic to hope that comprehensive IT management functions will ever be offered via any sort of common protocol. (See A Hopeless Situation.)

Bottom line? Sure, storage automation is an oxymoron of sorts. But one can certainly hope for more streamlining of storage management.