
The system uses a load balancer to shift tasks among servers based on their particular “cooling needs,” which are related to air pressure in the data center. The idea is to keep each server running as hot as they can without malfunctioning. The higher the temperature, the more you save on cooling costs.






