Improving power usage efficiency (PUE) in data centres

Can new technologies like liquid immersion cooling help greatly to improve data centres' power consumption?

Power usage in data centres is on the rise. The technological advances have given rise to the need for high-end computing infrastructure and subsequent data centre cooling and energy requirements. This is expected to continue increasing owing to expanded digitisation.

Businesses have a responsibility to promote data centres that are environmentally friendly in line with Environmental, Social, and Governance (ESG) principles.
So, how can data centres improve power usage efficiency? Businesses need to take measures to attain energy efficiency.

Power Usage Efficiency (PUE) is defined as the ratio of the power required to the consumption of total facility power. Total facility power represents the overall energy needs in running a data centre. It includes all the systems such as – servers, network facilities, storage, and elements such as cooling systems, batteries, fans, and generators among other equipment that are found in a data centre.

The following measures can help improve power usage efficiency in a data centre:

1. Locate the data centres in a relatively cool place and also where the power supply is stable

Cooling and power load factors have a direct bearing on power usage efficiency in a data centre. These two factors are significantly determined by the location of a data centre. Locating a data centre in a place where the ambient temperature is high means a relatively higher amount of power will be required to cool the infrastructure. Also, a constant power supply should be guaranteed to help in increasing efficiency. It is recommended to that if possible, data centres should be located in cooler areas of the country or city if power efficiency is to be achieved.

2. The general infrastructure

When thinking about the overall infrastructure of your data centre, look for self-sustaining solutions. The architecture should give room for natural light to enter the building while at the same time keeping the heat out.

3. Modify the data centre’s design

The designing and positioning of the racks in the data centre directly impact the power usage efficiency. Using a hot/cold model enhances efficiency without incurring huge investment costs. Minimal customisation of the data centres such as placing fans on the ceiling to allow sucking of hot air from the aisles will mean reduced power usage, helping in improving energy consumption.

4. Insulation

During hot or warm seasons, the temperature inside the data centre rises, necessitating cooling. Painting the data centre with a shield coating will significantly improve efficiency. Having vegetation on the rooftop not only helps in reducing the amount of carbon in the atmosphere but also serves as an insulator in the data centres.

You need to check the efficiency of the equipment in the data centre. If some equipment consumes a high amount of power segregate and keep them in a separate location.

Improving power usage efficiency in the data centre has numerous benefits including reducing costs as well as cutting carbon emissions.

How to Calculate your PUE 

PUE represents the ratio of the total amount of energy used by a computer data centre facility to the energy delivered to computing equipment:

PUE = Total energy entering the data centre / Energy used by IT equipment inside the data centre

 

Advanced and effective technologies such as liquid immersion cooling should be adopted to replace the legacy data centre design. With the use of liquid immersion cooling the above points are either greatly reduced or removed completely.  
The PUE of the data centres, like PeaSoup ECO Cloud, using liquid immersion cooling represents a real 1.01 mark or thereabout. The server immersed in dielectric liquid runs 50% – 70% cooler, and therefore requires much less power to operate and is expected to have a much longer, up to 40% operational, life span with less stress on the components, and long maintenance intervals, thus more reliable. 
Improving PUE in data centres