Asked about their cost containment strategies, 75% of U.S.-based data center managers selected server consolidation and 66% chose server virtualization. "These companies are looking to consolidate their data centers, such as 21 down to 14, as well as trying to load up the physical boxes with more virtual machines so they can get better utilization of what's already in their data center," says Matt Fairbanks, senior director of product marketing at Symantec. More than half (57%) chose "increasing automation of routine tasks" and 46% selected "cross-training IT staff." (The non-U.S.-based data center managers ranked automation of routine tasks slightly above server consolidation and virtualization.)

The survey turned up some alarming news about service level agreements: of the 65% of respondents who have formal SLAs in place, more than half (51%) have not met their SLAs in two years. Among all respondents, 85% said service level expectations have been increasing over the last two years. In addition, 31% of the managers said their data centers are not prepared for a major disaster. "If you look at the level of satisfaction of business with IT, it's generally not high," Fairbanks says. "There are many reasons for that, one is complexity continues to spiral out of control. Another problem is staffing – people write SLAs for the current staffing capacity and growth projections, but things change over 12 months."

Staffing issues uncovered in the study were quite shocking – 52% of managers said their data centers are understaffed. Finding qualified staff was a problem for 86% of these managers, retaining employees was an issue for 57%. Other problems cited were: skills do not meet current needs (57%), too many applications to manage (66%), staff skill sets too narrow (60%), and data centers too complex for staff (68%). "We found in focus groups that the skills required of the typical data center professional are not as deep and far more broad – they need somebody who can have a lot of balls in the air and juggle a lot of different disciplines as opposed to somebody with deep, specialized knowledge," says Fairbanks.