Maintaining appropriate data center airflow is critical for enabling the facility to handle temperature changes without damaging the equipment inside. The entire facility will become more efficient when air flows as it should. However, data center airflow management is a complicated matter requiring detailed assessments. Many people are also beginning to use advanced technologies for the best results.
Achieving Data Center Airflow Optimization With Artificial Intelligence
While artificial intelligence (AI) has upended and enhanced many processes, people are still exploring it to determine its full potential. Some have already had good results applying the technology to data center airflow management.
When a data center contains many pieces of hardware used for AI computing, the facility’s cooling needs increase. As computing becomes more advanced and powerful, people must adapt how they keep it at the right temperature to perform as expected.
AI Hardware Has Unique Data Center Airflow Management Needs
Tate Cantrell is the chief technology officer at Verne Global. He explained how data centers handling AI computing tasks have different cooling needs than other facilities. For example, they have higher-density server racks and need extremely low-latency network connections within the data center.
Cantrell clarified that four AI supercomputing systems in a single rack could use more than 40 kilowatts of energy while only consuming 60% of the space taken up by conventional data center racks. Racks containing AI hardware also get placed closer together to optimize bandwidth capacity and latency. If the racks use an air-cooled approach to cooling and lack enough space between them, complications can exert back pressure on cooling fans.
However, just as AI can add particular data center airflow needs to a facility, it can also address some of the challenges of keeping all the equipment at the proper temperature. For example, one company offers AI-powered software that monitors temperature changes throughout a facility and gives individual inputs to each unit to address any problems.
The solution can also evaluate the data center airflow impact of using one fan versus 50. This carefully tailored, data-driven approach can save up to 35% of the cooling power-related energy costs.
In addition to real-time data analysis that continuously monitors the environment to provide targeted cooling without wasting energy, AI technology checks for hot spots and other potential risks. It helps data center managers address these factors to avoid the consequences of poor data center airflow management.
Improving Data Center Airflow With Design Tools and Digital Twins
Designing a data center for optimal performance can be far more difficult than it initially seems. Even something as seemingly minor as good cable management can improve airflow around the facility. More specifically, poorly routed cables could block vents, making the associated equipment unable to dissipate heat through its internal fans.
Using a cloud-based data center design tool can help people from various locations collaborate, giving valuable input to ensure a newly built or remodeled facility functions as well as possible upon completion. Some cloud tools also have settings to enable clients to see data center plans and other progress but not edit documents. Such visibility helps keep everyone on the same page and sets the correct expectations.
Digital twins are also instrumental in achieving good results for data center airflow. They provide highly detailed models of real-life assets so people can experiment in a controlled environment before making decisions.
One company specializing in 3D digital twins for data centers offers models that show changing workload impacts over time. It lets users make alterations to improve data center airflow management and invest in the proper equipment or put it in the best locations to meet future needs.
The product relies on a science-based platform that calculates the past, present and future throughput associated with any chosen moment in time. It goes far beyond the pre-construction considerations a data center design team may make, allowing people to study the impacts specific choices have years into the future. Digital twins boost visibility and help juggle changing computing needs as technologies accelerate.
Applying Virtual and Reality to Design Efforts
Some people are also exploring how virtual reality could help them come up with more effective airflow strategies for data centers. One innovation developed by computer scientists at the University of California San Diego and Carnegie Mellon University combines real-world sensing technology with virtual reality.
While this handheld device is not specifically for data centers, it could assist in their oversight. The team designed it for maintenance personnel troubleshooting issues in any commercial building. In addition to measuring the environment’s airflow, it captures temperature and carbon dioxide details. The virtual reality component connects to the building’s electronic control system and shows the user relevant metadata and sensor information.
Once someone reports an issue affecting a specific part of a building, the technician can scan the area with a smartphone-based lidar tool, creating a virtual reality version of the environment. Data from specific sensors showing airflow measurements, temperature or other specifics are overlaid onto the simulated environment, which makes pinpointing and addressing problems easier.
The developers hope to build future versions of this tool geared towards occupants rather than just technicians. Then, even people without data center airflow expertise could notice problems and notify experts sooner. If such scenarios happen, data from scanners and sensors could be instrumental in data centers’ preventive maintenance strategies.
High-Tech Ways to Enhance Data Center Airflow
Succeeding with data center airflow management previously meant making detailed calculations while understanding the impact of aspects such as computing resources and the space between server racks. The above examples show how advanced technologies have an increasingly important role in all data center design aspects, including airflow management.
Add Comment