|
Greening Campus Data Centers
|
||
|
Seung-Jin Lee and Brendan O'Donnell
(sjlee81@u.washington.edu; bgo@u.washington.edu) |
||
|
|
Purpose
To recommend a highly energy-efficient design for the new Biology building's data center.
Benefits and Cost Savings to UW
Data centers are notorious for their high electricity consumption. Rising power densities and electricity prices fuel the need to implement energy-efficient designs in UW's data centers. The UW will need an estimated 2.5 to 3.0 MW of computing power in the coming decade. With the recent acquisition of the UW Tower Properties, three quarters of that power need can be met through the implementation of a new primary data center. The UW Tower data center has constraints that limit the most energy-efficient designs. By proposing a state-of-the-art sustainable design for a new data center, the university can save tens of thousands of dollars by simply reducing electricity consumption.
Analyses and Results
We can provide an order of magnitude cost savings based on a 0.4 MW server room facility. Assuming non-stop annual operation, a conservative energy savings of 10% is approximately equal to 350 MWh of electricity. With current energy prices of $0.07/kWh, a yearly savings of about $24,000 is possible. This is roughly equal to a graduate student's salary.
Implementation
With partial funding from Paul Allen, Ed Lazowska, the Bill & Melinda Gates Chair of Computer Science & Engineering at the UW, is heading a project for a new super-green biology building on campus. The project team involves faculty and graduate students from the College of Engineering, the College of Architecture and the Department of Biology. The goal of the project is to design a new 400,000 square foot biology building that will integrate large greenhouses, a highly efficient data center, ubiquitous sensors, advanced automation, etc., to achieve significant new levels of efficiency and to serve as a "living laboratory." This new building's proposed location is near Kincaid Hall, around the greenhouses on campus.
Indicators
Indicators for the proposed server room in the new biology building are straightforward and directly related to energy consumption. The efficiency of the server room can be measured in MW of electricity saved per year and ft^2 of floor space used when compared to traditional server rooms. We use these estimates to calculate an order of magnitude energy and cost savings.
Other Resources
|
|