Data center cooling method with critical device prioritization
A method and/or computer program product automatically adjusts room temperature in a data center room. A processor in a heating, ventilation and air conditioning (HVAC) system receives a throttle threshold temperature for each of multiple computing devices that are within a data center room, which is climate-controlled by the HVAC system. The processor in the HVAC system monitors a real-time temperature of at least one of the multiple computing devices. In response to the real-time temperature of said at least one of the multiple computing devices exceeding the throttle threshold temperature, an ambient air temperature in the data center room is decreased by adjusting a hardware thermostat in the HVAC system.
1. A method of automatically adjusting room temperature in a data center room, the method comprising:
receiving, by a processor in a heating, ventilation and air conditioning (HVAC) system, a throttle threshold temperature for each of multiple computing devices, wherein the multiple computing devices are physically located within a data center room that is climate-controlled by the HVAC system, wherein the throttle threshold temperature is a temperature that, if exceeded, causes operations of one or more components of a computing device to be reduced, and wherein the throttle threshold temperature is provided by a hardware management module (MM) associated with at least one of the multiple computing devices;
monitoring, by the processor in the HVAC system, a real-time temperature of at least one of the multiple computing devices, wherein the real-time temperature is provided by the hardware MM associated with said at least one of the multiple computing devices;
in response to the real-time temperature of said at least one of the multiple computing devices exceeding the throttle threshold temperature for said at least one of the multiple computing devices, decreasing an ambient air temperature in the data center room by adjusting a hardware thermostat in the HVAC system;
identifying, by the processor in the HVAC system, a critical computing device from the multiple computing devices, wherein the critical computing device has been predetermined to be critical to a mission;
identifying, by the processor in the HVAC system, a non-critical computing device from the multiple computing devices, wherein the non-critical computing device has been predetermined to be non-critical to the mission; and
prioritizing, for the mission, cooling of the critical computing device over the non-critical computing device, the prioritizing including controlling the ambient air temperature in the data center room based on the throttle threshold temperature for the critical computing device, or a throttling history of the critical computing device.
2. The method of claim 1 :
wherein the throttle threshold temperature of the critical computing device is a first temperature;
wherein the throttle threshold temperature of the non-critical computing device is a second temperature that is lower than the first temperature; and
wherein the prioritizing comprises selectively controlling the ambient air temperature in the data center room such that the throttle threshold temperature for the critical computing device is not reached and the throttle threshold temperature for the non-critical computing device is exceeded.
3. The method of claim 1 :
wherein the critical computing device is mounted within a first containment device;
wherein the non-critical computing device is mounted within a second containment device;
tracking the throttling history of the critical computing device; and
wherein the prioritizing comprises, in response to the throttling history of the critical computing device exceeding a predetermined frequency, adjusting airflow from the HVAC system to provide additional cooling air to a first area around the first containment device and a reduced amount of cooling air to a second area around the second containment device.
4. The method of claim 1 , wherein the multiple computing devices are all server chasses.
5. The method of claim 1 , wherein the multiple computing devices are all server blades.
6. The method of claim 1 , wherein the multiple computing devices are a combination of server chasses and individual servers.
7. The method of claim 1 , wherein the one or more components of a computing device comprise a processor, a memory, and a platform controller hub.