Data Center Efficiency: No Leader in x64 Server Market

IT shops don't perceive any single vendor as being more efficient than its competitors.

Which is the best x64 platform of them all?

A new survey from market-watcher Gabriel Consulting tries to distinguishes between the entries in the crowded x64 server segment but finds that, in many areas, customers don't perceive any single vendor as being a clear frontrunner.

X64 power, cooling, and efficiency are particularly tendentious subject areas.

It's been that way ever since the advent of the hyper-dense (and growing ever denser) data center. As industry veteran Dan Olds, a principal with Gabriel Consulting, points out, the larger server vendors have been "competing on the issues of energy efficiency, cooling, and data center density for half a decade now."

"[E]very vendor has made a major play to convince customers that their offerings are best at reducing power, heat, and footprint requirements," he notes.

In some cases, he explains, vendors don't appear to have made any progress.

"Customers really don't see a clear leader in terms of system energy efficiency," writes Olds, citing Gabriel Consulting's survey of 200 enterprise IT organizations.

For example, although a plurality of respondents (13 percent) perceive Hewlett-Packard Co. (HP) as being a leader in the development of energy efficient systems, no single vendor managed to crack the 15 percent threshold. HP finished ahead of rivals Oracle Corp. (at 10 percent), IBM Corp. (at 8 percent) and Dell Computer Corp. (at 7 percent). From the perspective of most IT shops, however, none of the big server vendors has effectively distinguished itself with respect to system efficiency. More than a quarter (28 percent) of respondents said they "weren't sure" which vendor had the edge in energy-efficient systems, while a full third said there was "No difference between vendors."

That makes a clear majority (61 percent) who can't distinguish a difference.

What Gabriel Consulting's survey might be reflecting is a kind of parity among vendors. "In past surveys, customers were significantly more decisive on this topic," Olds writes. "In 2009, 39 percent were in the 'No Difference' and 'Not Sure' camps. In 2007, only 27 percent of customers felt this way."

The same goes for system density -- i.e., the capacity to cram N number of systems into X cubic feet of data center space. HP again led the way, followed this time by IBM, but -- as was the case with system efficiency -- the "nays" (i.e., the "Not sures" and "No differences") outweighed the others, garnering 58 percent of all responses.

Is this a trend? In all but two functional areas, a plurality of respondents named HP No. 1 -- while a majority cited "Not sure" or "No difference." In two additional functional areas (viz., "Best Data Center Design/Advisory Service" and "Vendor Best Able to Increase IT Efficiency"), IBM and HP traded places -- but still finished a distant second behind the "nays."

If IT consumers are frustrated, vendors are probably doubly so, Olds suggests.

"Given how much effort the systems houses have devoted to revamping their product lines to reduce power usage and increase space utilization, I'm sure they're hoping for a better mindshare return on those investments. However, it's difficult to make a name for yourself in a certain area when every other vendor is trying to stake out the exact same real estate," he observes. That said, the stakes in the x64 power and cooling race are about to change.

In this respect, Olds suggests, vendors should have a golden opportunity to distinguish themselves. "We see facilities efficiency becoming even more important in the future. The ties between data centers and their facilities management are becoming tighter," he writes. "Power, cooling, and floor space consumption are [important factors] in purchasing decisions now and will become more important in the future, according to our survey respondents."

About the Author

Stephen Swoyer is a Nashville, TN-based freelance journalist who writes about technology.