Data Centers: Poor Air Distribution Leads to Hot Spots
With the rapid development of new, more powerful information-technology (IT) devices in institutional and commercial buildings, the need for expanded cooling capabilities in data centers is becoming a higher priority for many facility managers.
Data centers contain various types of equipment, such as uninterruptible power supplies (UPS), firewalls, routers, switches, and other IT devices. With these additional pieces of equipment come an increased heat load and hot spots — potential problems for which managers must account.
But how do managers tackle the cooling challenge? What are the major factors involved in cooling these spaces, and what steps can managers take to protect critical equipment and operations?
About one in ten servers runs hotter than published reliability guidelines, according to the Uptime Institute, a research organization focusing on improving data-center operations. Hot spots occur often in data centers and can result from under-designed airflow distribution in data centers with under-floor plenums. Common problems include bypass flow, recirculation flow, and negative flow.
Bypass flow occurs when cool air from a computer-room air-conditioning unit returns to the return-air plenum before it passes through or over the data-center equipment.
Recirculation flow occurs when air that has passed through or over the equipment does not reach the return-air plenum but instead remains in the space to pass over the equipment again.
Negative, or venturi, flow occurs when the supply-air velocity is too high, and room air enters the supply plenum, increasing the temperature of the air delivered to the data-center equipment.