Advanced cooling concepts for data centers
The successful operation of a large-scale air-cooled data center (DC) requires an efficient cooling system to ensure that the DC operator can provide its services to customers with maximum availability and reliability at minimal operational cost. An efficient cooling system guarantees that the temperatures at the inlets of all computing devices in the DC racks never exceed a given threshold value. This prevents device overheating and achieves this goal with minimum required cooling energy. However, most of today's DCs waste cooling energy because they are operated at significantly lower temperatures than actually necessary. This approach certainly reduces the potential risk of reacting too late to a harmful temperature increase in the DC, but can lead to the consumption of 4 to 6% more cooling energy for each degree Celsius below the upper temperature limit as we have shown with CFD simulations and a detailed cooling system model.
To realize these potential savings in cooling energy, we are working on the design of advanced sensing and control concepts to optimize the air flows and the temperature distribution in the DC room. Together with partners from the Zurich University of Applied Sciences in Winterthur and Siemens Building Technologies in Zug, we have developed a robust air-flow control system that uses variable air-volumn actuators to adjust dynamically the cooling air flow from the computer room air conditioning (CRAC) unit to the server racks so that all server inlet temperature values are close to, but never above, a maximum threshold value. We verified the implemented control concept by carrying out experiments in a test room of the IBM DC in Boeblingen with an open and closed control loop for various symmetric and asymmetric server loads. The results demonstrate that up to 20% of cooling energy can be saved in the test room by applying dynamic air-flow control. Using the proposed control concept in a large-scale DC and incorporating the control strategy into a WSAN are issues that have to be studied next.
