Optimal temperature and airflow

Artificial intelligence (AI) is changing the technology landscape in various industries, and data centers are no exception. AI algorithms are computationally heavy and will increase data centers’ power consumption and cooling requirements. This aspect is arguably the one that will most deeply affect data center architecture. That being said, AI will also help automate some aspects of data center operation and maintenance.

This article will elaborate on how these AI aspects will change data center architecture.

Energy Efficiency

Data centers famously consume a significant amount of energy, and reducing their power consumption is an essential goal for data center operators.

The Uptime Institute estimates that the average power usage effectiveness (PUE) ratio for data centers in 2022 is 1.55. This implies that for every 1 kWh used to power data center equipment, an extra 0.55 kWh—about 35% of total power consumption—is needed to power auxiliary equipment like lighting and, more importantly, cooling. The Uptime Institute has only observed a marginal improvement of 10% in the average data center PUE since 2014 (which was 1.7 back then).

In some ways, AI can help data centers to become more energy efficient. By analyzing temperature data, heat generation, and other variables, AI can determine the optimal temperature and airflow for different areas of the data center. By looking at energy consumption and heat generation data, AI can determine the optimal placement of equipment to minimize energy consumption and reduce wasted heat. By studying historical data and forecasting future energy consumption, AI can help data center operators to identify areas where energy consumption can be reduced.

While these benefits have the potential to reduce data center power consumption in the future significantly, the current reality is that power-hungry AI algorithms require more computing resources and power and will lead to a net increase in data center power consumption. The world’s major data center providers are already gearing up for this increase.

For example, a recent Reuters report explains how Meta computing clusters needed 24 to 32 times the networking capacity. This increase required a redesign of the clusters and data centers to include new liquid cooling systems.

Intelligent Automation

AI can automate and optimize many monitoring and scheduling tasks currently performed manually in data centers. These optimization and automation processes could also reduce the amount of hardware to purchase, manage, and monitor, as explained by Pratik Gupta, CTO at IBM Automation.

For example, AI can be used to automate capacity planning. By analyzing historical data and forecasting future demand, AI allows data center operators to determine the optimal computing resources to support current and future workloads. Such work makes planning for growth easier and ensures that data centers have sufficient resources to support their customers’ needs.

AI for Self-Healing Data Centers

A VentureBeat report explains that there has been a trend in using AI for fault detection and prediction in data centers in recent years. This leads to “self-healing” mechanisms that help data center operators reduce downtime and improve the reliability of their infrastructure.

For example, AIs can help monitor traffic throughout the data center. If traffic in specific nodes is slowing down, the AI can detect that trend and find solutions to restart the node or reroute traffic to other nodes. These trends and issues might not be immediately apparent to human operators.

With AI, data centers can look at historical data to predict equipment failures and schedule maintenance before a failure occurs. This can help prevent downtime and ensure that equipment always operates at peak performance.

Takeaways

AI can significantly improve the operation of data centers by automating and optimizing resource allocation, as well as doing predictive maintenance and fault detection. These processes can help data centers become more energy efficient in the long term, but in the short term, AI will lead to significant increases in data center consumption. This is shown by how major data center providers have had to restructure their data centers to include enhanced cooling capabilities, such as liquid cooling.

Tags: Artificial intelligence (AI), Cooling requirements, Data center architecture, data centers, energy efficiency, Intelligent automation, Networking capacity, Optimal temperature and airflow, power consumption, Power usage effectiveness (PUE) ratio