PUE
What is it?
PUE (Power Usage Effectiveness) is the ratio of the total amount of power used by a data center compared to the power that's actually delivered to the computing equipment. An ideal PUE ratio would be 1.0: 100% of the power delivered to the data center is used for computing. However, data centers require cooling systems, lighting, and other equipment that will also consume some of this power. As such, the PUE for a conventional data center will always be greater than 1.0.
Why do you need it?
In a large data center, operational expenses (such as power bills) can often be enormous and a heavy burden that persists after the initial cost of building the data center and buying the equipment. To reduce expenses, companies strive to reduce both the energy consumption of the servers and the power that's used to maintain and cool the equipment.
Traditional cooling infrastructure such as air conditioning systems use a lot of energy. Therefore, companies are looking to reduce the PUE of their data centers in many different ways – such as using closed loop liquid cooling systems (instead of fan cooling) on the servers themselves, which are more efficient at cooling the servers, and therefore allow the servers to operate at higher ambient temperatures (requiring less or no additional air conditioning); or even using liquid immersion cooling where the entire server rack is immersed in a non-conductive fluid for cooling.
Other companies can reduce their PUE by moving their data centers to a colder climate – such as northern Europe, Russia, or Alaska, requiring less additional energy to reach a lower ambient temperature. According to reports, Microsoft is even experimenting with building data centers under the ocean.
How is GIGABYTE helpful?
GIGABYTE has formed close partnerships with vendors of both closed loop liquid cooling systems as well as immersion cooling systems to offer a variety of different solutions to help customers reduce their PUE. Partnering with these companies allows GIGABYTE to offer our servers that have been modified, tested, and validated to work with these different solutions to provide efficient and effective deployment.
