Data-processing and storage demands have soared over the years as organizations and customers have demanded quicker access to a larger array of information and resources, from accessing sales reports from past fiscal years to conducting real-time teleconferences to viewing and sharing high-def videos of Aunt Linda's new baby hiccupping for three minutes straight. Plenty of datacenter operators have embraced perhaps the simplest solution possible: throwing hardware at the problem.
Only later do problems with this approach begin to surface. For example, datacenter operators have failed to accommodate for the effect that, say, doubling the IT load will have on cooling and airflow. In a vain effort to keep all the machines operating at a safe temperature, datacenter operators have turned to cranking up CRAC units to the max -- or to put it another way, blowing cold air at the problem. This has proven costly as many datacenters operator find themselves paying as much to cool IT hardware as they pay to power it.
[ Learn how datacenter operator RagingWire grapples with cooling challenges using SynapSense's wireless monitoring system. | Utilities across the country are offering incentives to companies that cut energy consumption. ]
Fortunately, IT companies such as IBM, AdaptivCool, SynapSense, and others are rolling out products and services to assist datacenter operators in managing cooling and airflow issues. Doing so can help an organization reclaim a hefty chunk of its power and utilty budget, often without having to invest time and money into ripping and replacing existing equipment.
Measure to manage
IBM, for example, announced this week that Toyota Motor Sales successfully deployed its Measurement and Management Technologies (MMT) over a five-month period at Toyota's 20,000-square-foot datacenter in Torrance, Calif. Big Blue's multilevel measurement tool assesses thermal readings throughout a datacenter -- from floor to ceiling -- and provides a detailed breakdown of the heat distribution by creating a high-resolution, three-dimensional chart that pinpoints power and cooling inefficiencies.
According to Hendrik Hamman, research staff member and manager of the Photonics and Thermal Physics department at IBM, those inefficiencies might relate to the physical placement of servers and other IT gear, as well as cooling ducts and even perforated tiles. The map might also reveal which CRAC units can be turned off entirely.