Energy consumption occupies great bulks in the cost of operating a data center. Thus energy efficiency is always the main problem for every organization. Leading tech companies like Apple, Google and Microsoft have already unlocked the door towards more energy efficient data centers, but many other organizations still operate their servers with energy-devouring systems.
There are best practices that data center operators can implement in order to cut cost. And for owners, investment on highly-efficient subsystems is important.
This quick list give you a glimpse of the best practices that you can implement for a more energy efficient data center.
Data Center Best Practices
Cooling System (Use Free Cooling)
Even when the temperature outside is cold, data centers require cooling every second of the day. Cooling alone is a monstrous energy-devourer. Data centers that utilize chillers or air-conditioning units eat up 30-70% of non-computing energy.
Thanks to our Mother Earth, even data centers that need a reliable cooling system can be supported by natural cooling agents found in the environment. Using natural free cooling is one of the best practices for more efficient data centers.
Outside air can provide “free-cooling” cycles to data centers and helps reduce chiller or compressor operation in precision cooling units. According to Emerson Network Power, using natural air flow can provide energy savings of 30-50% depending on the site.
Water-side economizer is also a natural solution to save energy use. This economizer system uses the evaporative cooling capacity of a cooling tower or dry cooler to satisfy cooling requirement during mild outdoor conditions especially during the night in sites with hot climate.
Match Cooling Capacity and Airflow with IT Loads
“The most efficient cooling system is one that matches needs to requirements.” This is what Emerson Network Power proposes as part of data center best practices.
The control system helps in meeting this end by working like an octopus with multiple cooling units as tentacles. Control system allows to determine what action needs to be undertaken in order to maintain the “conditions in the room at target levels and match airflow to the load.”
The whitepaper advises that an intelligent cooling control is the key solution in order to meet the cooling requirements of IT loads. An intelligent cooling control is one that has the capacity to understand, forecast and change the cooling capacity and airflow based on the conditions at the server, and NOT based on return air temperature.
Moreover, the 70°F temperature rule for data centers is now relegated as a myth. Data center operators now can raise the temperature of the cold aisle up to 10°F higher. And this increase is equal to a 20% reduction to energy used in cooling system.
Optimizing Power Distribution with Highly Efficient UPS
One of the core concerns in operating a data center is that it should function uninterruptedly. And the Uninterruptible Power Supply or UPS is the one in charge for it. However, a portion of the energy supplied by UPS to the IT equipment is wasted as heat. This occupies a large portion in the losses incurred in the power distribution. That is why, it is imperative to choose a high-efficiency model UPS.
The best practice to do is to choose the UPS system with the highest efficiency that can meet the power conditioning needs of the data center. An annual savings of over $38,000 per 0.14 hectare-data center can be accumulated by merely raising the efficiency to 5% according to PG & E.
By improving the efficiency of the UPS system, data centers can benefit an hourly energy savings within the UPS itself alone.
Increase Visibility, Control and Efficiency with DCIM
Data Center Infrastructure Management or what is called DCIM is a great tool for translating monitoring data into a manageable and useful information for decision making and creating solutions.
With the use of DCIM, all subsystems of the data center e.g the power system, cooling system, and servers are all optimized holistically providing real-time visibility and control.
DCIM can automatically extract real-time energy usage, accurately display overall trending information, and zoom into the details such as the fluctuation of temperature. With this, data center operators can troubleshoot problems and prevent impending ones with the holistic view of the data center.
Bonus: What Can Other Data Centers Learn from Google?
Make Comprehensive Measurement of Power Usage Efficiency
While the internet use (and google search) continues to grow tremendously, the burden to keep the site running uninterruptedly to its best capacities is a hard and arduous to manage.
Google reduced the energy use of its data centers by cutting down overhead energy to only 12% and focusing more on the energy that powers the server. Thus Google services is able to keep pace with the growing number of internet users.
Comprehensive PUE measure. Google boasts PUE of 1.12 over 12-month period of trailing twelve-month (TTM) in all of its data centers, in all seasons while the average PUE of other data centers is around 1.7.
Thus, a comprehensive measurement of PUE is important in order to point out what needs to be addressed. A comprehensive PUE monitoring is important in order to reduce overhead power, or energy allocated to non-computing functions like cooling and power distribution.
As Google said: “You can’t manage what you don’t measure, so be sure to track your data center’s energy use.”
About the Author, Micah de Jesus
Micah de Jesus is a Digital Marketing Professional and a data / cybersecurity news junkie. She works as the Managing Director of GrowthScout SEO, a digital marketing firm and a Contributing Editor for INOC – Network Monitoring Center, The Partnership UK, Turrem Data and Solus Multi Factor Authentication.