Modern data centers are massive consumers of energy. They also remain surprisingly energy inefficient. But carefully match their energy supply with demand as dictated by their workloads, and you can seriously reduce those inefficiencies.
Combine that with holistic energy demand techniques, and you can lower total data center power costs by 30% and dependence on a larger power grid by more than 80% – plus cut your capital and operational costs as well.
That’s the argument presented by researchers from HP Labs’ Sustainable Ecosystems Research Group in a paper this week at the IEEE’s 13th annual Intersociety Conference on Thermal and Thermomechanical Phenomena in Electrical Systems.
Titled “Towards the Design and Operation of Net-Zero Energy Data Centers,” the paper outlines an innovative management architecture that integrates energy and cooling resources with IT workload planning. It saves energy by employing modules that:
- Leverage powerful predictive analytics software to forecast the availability and cost of critical resources, like renewable energy, and IT workload demand.
- Employ a novel optimization algorithm to schedule workloads based on resource availability.
- Manage workload and energy consumption in real-time according to performance requirements and data center operational objectives.
- Use verification and reporting to ensure plan accuracy.
Design data centers in this way and they can be run by stand-alone micro-grids powered by renewable resources such as sunlight, wind or water – resulting in a net-zero drain on the larger power grid and substantial savings in operating costs for data center owners.
“Information technology has the power to be an equalizer across societies, but the cost of IT services, and by extension the cost of energy, is prohibitive to widespread adoption,” noted Cullen Bash, Distinguished Technologist, and interim director of HP’s Sustainable Ecosystems Research Group in announcing the new architecture. “The HP Net-Zero Energy Data Center not only aims to minimize the environmental impact of computing, but it also has a goal to reduce energy costs associated with global data center operations to extend the reach of IT globally.”
By reducing cost of data center ownership, Bash hopes, organizations of all sizes will be able to take advantage of the power of IT, while tailoring the underlying technology that they use to meet their specific business needs.