The University’s Estates Team commissioned Future-tech to carry out a Data Centre Optimisation Survey. This survey assessed the physical infrastructure that supports the data centre and provides a controlled environment for the servers and other IT equipment to operate in. The survey looked at three distinct areas within the data centre; power provision, environmental control and future expansion. The survey also highlighted current energy use, areas of inefficiency and elevated risk.
CHALLENGES AND CONSTRAINTS
Although the data centre was performing well, changes in technology and best practice presented areas that could offer improvement with regard to the data centre’s energy efficiency and air flow. When the data centre was originally designed the application of high density blade servers was not on the horizon. When the University’s IT Team invested in this new technology it reduced the overall number of servers required, thus reducing energy consumption. However the new high density servers concentrated more heat load into a smaller area, resulting in hot spots in certain parts of the facility. As there was no air segregation within the data hall itself, supply and return air was mixing which reduced the potential return air temperatures within the system. These lower temperatures returning back to the air handling units resulted in the system operating at a less efficient level than possible.
THE FUTURE-TECH SOLUTION
Removing hot spots and increasing return air temperatures can both be achieved by introducing an aisle containment system. Aisle containment segregates the supply and return air plenums; this means cooled air can only return to the air handling units after travelling through a server. This increases energy efficiency as all the cooled air has to perform “work” before returning to the air handling units. Aisle containment is only part of the solution as the new air flow paths and increased return air temperatures means the entire system needs to be recalibrated and commissioned to ensure the very best performance can be achieved by the cooling system.
Before the improvements were implanted PUE was 2.0, but then by increasing the return air temperatures and commissioning the system to its new parameters the data centre’s cooling system was improved by around 50%, taking the PUE to 1.5 or better. In real terms this translates to an energy saving of approximately £60,000 per year, giving the project a payback period of less than 12 months, which will also reduce the University’s carbon footprint by approximately 315 tones of CO2 per year.