Is server virtualization the answer to utilization and IT budgetary problems that it is widely purported to be? No, says prominent blogger, consultant and 26-year DR planning veteran John W. Tiogo, founder of Tiogo Partners International.

“The preponderance of virtualized systems are low-traffic Web servers and file servers that don’t need virtualization at all,” he argues. “Many companies are abandoning their server virtualization projects when they are about 20% complete, which is good news. I don’t trust virtualization with my most mission-critical applications, and neither do my clients.”

Virtualization, he said in an interview in the SiliconAngle Cube from IBM Edge 2012, is often used to solve a problem that no longer exists, or should not exist (full video below). “People are spending a ton of money to virtualize dinosaur technology.”

The issue is that IT keeps too many applications and too much data in memory on servers, waiting to be used. A decade ago, when CPUs had much less power and storage systems were slower, loading everything into memory at the start of the day was necessary. That became standard practice and is still followed today, with the result that IT overuses and wastes resources.

“I can push a button on my smartphone and start an app,” he said. “that app was loaded from storage somewhere.” If IT followed that model it would not need virtualization to conserve resources.

Value of Big Data

Tiogo also questions whether most large companies need big data resources in-house or would be better advised to turn big data questions over to a service provider, provided that effective big data service providers evolve. He agrees that big data research can be valuable for answering specific questions. The basic question is how often do these questions arise for the average enterprise.

“I don’t have the money for petaflops of processor and petabytes of storage just on the chance that I might want to ask more than one question,” he says. “Looking for voter fraud is not something I need to do every day, for instance. I can do it once every three years.”

He implies that vendors may be using big data as an excuse to sell huge amounts of hardware that will stand underused in company data centers. “The question is do I ask my Global 2000 clients to buy a lot of infrastructure to ask these questions, or are they once-in-a-blue-moon questions that they are better off farming off to a service provider.”