Welcome to Keep Your Cool - a series tackling simple cooling optimization strategies for the busy data center operators by former busy data center operator, Gregg Haley.

A couple of months ago, I wrote a blog called “A Baby Boomer’s first try at ChatGPT” that talked about my first experience using Chat GPT. Recently, I started wondering how I could apply Chat GPT to our workplace.

We’ve recently been working on a number of projects where we have been assessing the effectiveness of cold aisle containment. During these assessment, we typically generate a Static Heat Map, that shows the temperature distribution within the aisle using data collected by the AUDIT-BUDDY system at three heights (72", 42", and 12") from each cabinet. The color scale on the maps represents temperature, ranging from dark blue for colder areas to red for hotter areas, with multiple colors indicating potential airflow inefficiencies. The main goal for our clients is to achieve a more consistent temperature and color distribution across the map.

We became intrigued by the absolute uniformity of temperatures in some instances, while others recorded less than optimum performance. How could we rate performance and communicate the results to clients, so that they might take steps to address the poorer performers.

Statistically, the Standard Deviation is the best measurement of uniformity. The lower the standard deviation, the less variance in temperatures in a sample set. Having observed first hand the causes of a higher standard deviation in a sample set, I thought it might be interesting to ask ChatGPT what it would reply with. I first set about constructing my query to make sure it was focused sufficiently to generate the type of response I was expecting. A loosely structured query will result in a loosely structured response, so I wanted to make sure to hone in on the subject in question.

Query: In a cold aisle containment installation, in a data center, how important is uniformity of temperature across the vertical plane?

In a cold aisle containment installation in a data center, maintaining uniformity of temperature across the vertical plane is highly important. Here's why:

1. Equipment Performance: Data center equipment, such as servers and networking devices, operate optimally within a specified temperature range. Deviations from this range can lead to performance degradation, increased power consumption, and potential hardware failures. Uniform temperature distribution ensures that all equipment within the data center operates within the recommended temperature range, promoting efficient and reliable performance.

2. Hot Spot Prevention: Non-uniform temperature distribution can result in localized hot spots within the data center. Hot spots occur when certain areas experience significantly higher temperatures than others. These hot spots can lead to thermal stress on equipment, shortening their lifespan and increasing the risk of unexpected failures. Uniform temperature distribution helps prevent the formation of hot spots and maintains a balanced thermal environment throughout the data center.

3. Airflow Efficiency: In a cold aisle containment system, cool air is supplied from the cold aisles and hot air is exhausted from the hot aisles. Maintaining uniform temperature across the vertical plane ensures efficient airflow throughout the data center. It helps prevent the mixing of hot and cold air, minimizes recirculation, and enables effective cooling of equipment. A balanced airflow pattern enhances the cooling efficiency and reduces the workload on cooling systems, improving energy efficiency in the data center.

4. Predictability and Manageability: Uniform temperature distribution provides predictability and manageability in the data center environment. It allows for better control and monitoring of the cooling infrastructure. By ensuring consistent temperatures across the vertical plane, data center operators can implement temperature monitoring systems and cooling strategies more effectively. It also simplifies capacity planning, as the thermal conditions are more predictable and manageable when the temperature is uniformly distributed.

Overall, uniform temperature distribution across the vertical plane in a cold aisle containment installation is crucial for equipment performance, hot spot prevention, airflow efficiency, and overall manageability of the data center environment. It helps ensure reliable operation, optimal performance, and efficient cooling, ultimately contributing to the reliability and longevity of the data center infrastructure.

Purkay Labs is supplying this metric, when assessing the thermal performance of data centers, as an additional data point for our clients so they can compare the performance of their contained spaces against one another. We have also established some guidelines for interpreting the Metric. 

The Uniformity Metric is a measurement  of the temperature differentials, across the vertical plane of the cold aisle, when measured at the 12”, 42” and 72” heights for every rack in the aisle. The lower the number the more uniform the temperatures are across the entire vertical plane.

Uniformity rating of: 0 to 0.49 as VERY GOOD

Uniformity rating of: 0.5 to 0.99 as GOOD

Uniformity rating of: 1.0 to 1.49 as FAIR

Uniformity rating of: ⥸1.5 as NEEDS IMPROVEMENT

In a case study, we had sampled in a very tightly contained environment. It consisted of three rooms, each with 100 KW IT load, distributed evenly across 20 racks. The space was provisioned with containment, floor to ceiling, for both the hot and cold aisles. Each of the rooms calculated a Uniformity Metric of in the range of .44 to .49 which puts all three in the very good range. Other sites we have measured calculate a wide range of Uniformity from .437 to 4.03 for  different Cold Aisle Contained spaces. It must be noted that there are uncontained spaces, with a higher temperature, that exhibited a more uniform air distribution across the plane, resulting in calculations in the good to very good ranges.           

The takeaway from this is the higher the Uniformity Metric the more likely there are  influencers - like missing blanking plates, devices exhausting into the cold aisle, excessive CFM of cold air from a specific perforated tile, air leaking under the cabinet due to gaps created by cabinet feet, which are impacting the even distribution of cold air to the server inlets. 

About the Author

Gregg Haley is a data center and telecommunications executive with more than 30 years of leadership experience. Most recently served as the Senior Director of Data Center Operations - Global for Limelight Networks. Gregg provides data center assessment and optimization reviews showing businesses how to reduce operating expenses by identifying energy conservation opportunities. Through infrastructure optimization energy expenses can be reduced by 10% to 30%.

In addition to Gregg's data center efforts, he has a certification from the Disaster Recovery Institute International (DRII) as Business Continuity Planner. In November of 2005, Gregg was a founding member and Treasurer of the Association of Contingency Planners - Greater Boston Chapter, a non-profit industry association dedicated to the promotion and education of Business Continuity Planning. Gregg had served on the chapter's Board of Directors for the first four years. Gregg is also a past member of the American Society of Industrial Security (ASIS). Gregg currently serves as the Principal Consultant for Purkay Labs.