Data center operators take a variety of actions to eliminate hot spots. Some of these actions result in short term fixes but may come with an energy penalty and some actions may even create more hot spots. Airflow management is an easy and cost effective way to permanently eliminate hot spots while saving energy and can avoid the capital expense of adding more cooling units. This paper describes the root cause of hot spots, recommends methods to identify them, reviews the typical actions taken, and provides the best practices to eliminate them.
Hot spots can result in poor server reliability, performance, or even server damage. Hot spots typically appear at the intake of IT equipment due to inefficient air flow management like cold air leakage (i.e. bypass air) and hot exhaust air recirculation. Feeling it by walking around, taking manual measurements, or having automatic monitoring are three main ways to detect them. Many actions taken by data center operators to fix hot spots are either temporary emergency use only, ineffective, likely not required, or make matters worse. Rack and room airflow management, containment, relocating problem equipment, changing the location of the air temperature sensors, and allowing DCIM to control airflow of cooling units are the best practices which can help fix hot spots in an easier and cost-effective manner.