This isn't a situation where technology is being pushed through vendor hype. The ability to dramatically diminish server sprawl and its associated costs is so significant that organizations are demanding it -- not just for areas such as development, quality assurance and testing, but for full production as well.
As a storage guy, I must ask the question, Why hasn't storage virtualization taken off in the same way? The problems associated with server and storage management are remarkably similar: lots of devices to manage, poor resource utilization, and high consumption of data center power, cooling and floor space. So clearly the need exists.
For a technology to take hold, it must also:
-- Be conceptually easy to understand.
-- Demonstrate sufficient maturity and robustness to meet production standards.
-- Offer a clear payback.
It also doesn't hurt to have a loyal and vocal chorus of successful early adopters.
While server virtualization meets all of these criteria, storage virtualization unfortunately doesn't -- at least not yet. While the potential value is easy to understand, technical maturity and the potential for savings remain cloudy.
Just to be clear, while there is lots of virtualization within arrays, I'm referring to storage virtualization at the network layer. A handful of products have evolved over several generations and developed market niches providing specific functions, most notably data migration and network-attached storage global namespaces. Others still mainly offer future promises.
Depending on the environment, virtualization can either reduce or add to the level of management complexity. Concern with finger-pointing and vendor nonsupport of virtualized configurations is another problem. Confusion over competing technology approaches, lack of standards and fear of vendor lock-in are additional deterrents.
Server virtualization has succeeded because, relatively speaking, it is less disruptive. It doesn't require hardware replacement or new operating systems, and unlike storage, it doesn't require a massive data-migration effort to undo if it doesn't work out. Most importantly, it's largely transparent to applications.
I do believe storage virtualization will get there someday. Unfortunately for many environments, someday is not yet here.
Jim Damoulakis is chief technology officer of GlassHouse Technologies Inc., a leading provider of independent storage services. He can be reached at jimd@glasshouse.com.