Coming up with new approaches to battling the rising cost of operating enterprise-class data centers will require effort from throughout the IT industry, as well as from outside sources such as Carnegie Mellon University in Pittsburgh. The university this week opened a new data center that will provide computing resources to the institution and its students and serve as a research lab that will target issues related to rising energy costs and, perhaps even more importantly, management costs.InformationWeek on Monday will explore the ongoing efforts at Carnegie Mellon, as well as look at the issues surrounding escalating data center operational costs. Greg Ganger, professor and director of Carnegie Mellon's Parallel Data Lab, says utility bills to power and cool data centers are now greater than the cost to buy the equipment housed. But that's not the big-ticket item. Ganger says the university's research with dozens of IT executives indicates that the human administration costs are now four to seven times greater than capital expenditures.

"If you look at operational costs, we've recently crossed the line where power and cooling is more expensive than the equipment," Ganger says. "That other half of the equation, human administration, has long been bigger than the cost of equipment. Those lines crossed years ago. We've got to try and understand where all that human time is going. It is amazing how poorly as an industry that we understand these costs."

Carnegie Mellon this week officially opened the Data Center Observatory, a 2,000-sq. ft. data center that will test and help refine many of the newest technologies and equipment available for improving energy efficiency. The center has also begun the task of cataloging exactly what tasks and jobs within its data center workers are engaged in during the setup and operation of the facility. Within six months, the university hopes to use the information to begin developing strategies for attacking the problem of management costs.

Carnegie Mellon will be just one element needed within the industry to begin addressing these mounting data center operational costs, whether associated with energy or human resources. Most major technology companies have ongoing efforts to create more energy-efficient processors, systems, and data center cooling equipment, and many are engaged in projects to improve data center automation. An example of consortium-level efforts include the recently formed The Green Grid, which now includes Advanced Micro Devices, Dell, Hewlett-Packard, IBM, and Sun.

Ganger says neither Carnegie Mellon by itself, nor any other single university or industry consortium, will be able to come up with all the answers.

"This is a long endeavor, and it will involve a lot of people--an army of researchers," Ganger says.