Data centers consume the most energy per square foot than any other use. The energy goes to powering the racks and hardware and cooling. Further data centers form the most protected, mission critical territory in any company. All of the data is on the servers: from payroll to accounts receivable to corporate work product to lunch menus.
To reduce the energy consumption, you can virtualize your servers if you have the budget or try to upgrade to more efficient equipment. The only other place to look for energy savings is the cooling. That is where Kilawatt Technologies has concentrated and succeeded.
The following articles provides data we have collected from actual data centers and ways we have significantly reduced the consumption.
When we start discussing data centers with businesses, one of the first objections is that the environmental conditions are too strict, hence outside air cooling, for example, can not be accomplished.
The primary parameter that is raised is Relative Humidity. If theRelative Humidity is to low, then static issues will develop. The range of RH that is "required" is reported to be from 35% to 50%.
We decided to find out if this was actually true. We went into a data center and instrumented the racks to determine exactly what relative humidity they were seeing. The data center had a humidity control system that they thought was maintaining a minimum of 40% RH within the data center. There was no outside air cooling operational in the data center. The IT department identified no known issues and even made statements that they thought the RH was within acceptable levels.
The results: The cold side of the rack was seeing RH that averaged 11.5%, ranging from 5.2% to 24.4%!
With proper grounding, which is a common sense requirement, Relative Humidity is not a factor.
But, please do not take just our word for it. Intel has performed a test using 100% outside air, no humidity control, limited filtering of air for 10 months and reported no issues with data center operation. The Relative Humidity ranged from 5% to 30% during the test period.
What cooling power is a rack of computer equipment actually seeing?
Is the cooling system you have actually providing efficent cooling?
Both good questions. What we found was surprising to both us and a corporate IT department. The Data Center we worked had a raised floor diffuser system supplied by a stand along, dedicated mechanical cooling system. This unit provided cool air 24/7 in an effort to maintain the rack at the "manufacturing required" operating temperature.
We instrumented the cooing rack with temperatureprobes for ten days. We compared the temperatures coming from the diffusers to that entering the cold side of the rack and the exhaust out of the hot side of the rack.
Temperature at the diffuser averaged 54.6 Degrees F
Temperature at cool side of the rack averaged 75.7 degrees F
Temperature at hot side of the rack averaged 88.6 degrees F
The air temperature increased MORE from the time it left the diffuser than it did crossing the rack!
Of course this leads to the question: How big of an issue is this? The IT department and Kilawatt reviewed the manufactures specifications and realized that the manufacturer has specified a maximum operating temperature of 70 degrees C. Therefore, the system was within specification.