Steven Carlini, Vice President of Innovation and Data Center, Energy Management Business Unit, Schneider Electric.
It is satisfying when you make predictions and they come true because that’s not always the case. In a post I wrote six years ago, I predicted the data center industry would leverage AI, and I am happy to report that I was correct!
If you are operating a data center, you can leverage AI today for meaningful benefits in several ways. I am not talking about content generation AI. I am talking about trained, language learning models for autonomous decision-making, process and system automation, forecasting, anomaly detection, and prescriptive analytics. Let’s dive into the specifics.
AI Applications In Data Centers
Here are some AI uses and their applications in data centers.
Weather Prediction
Cloud-based trained AI models offer accurate, real-time forecasting for all of your data center sites, enabling you to migrate workloads across your data center architecture to improve resiliency.
Utility Power Control
Leveraging AI can help data center operators by predicting and automating utility power sources. Operators can have multiple utility feeds for core utilities and also distributed utility sources. Operators can use an advanced distribution management system (ADMS) to specify preferences around the lowest carbon source or lowest cost source, for example. Then, the AI model can predict which sources will be available and automate the power selection.
Back-Up Power Control
Utility grids are much less stable than a few years ago. Many data center operators are not only using UPS and minimal batteries with generators, but they are also moving to multi-hour battery storage onsite. This allows them to peak shave when utilities use “high demand pricing,” fast frequency response when grids are unstable, charge the batteries at optimal cost or carbon footprint, and even monetize by selling energy to utilities when they need it.
Back-up power control is going to get even more complicated and data center operators are risk averse. AI can help predict and automate these functions while ensuring proper resiliency.
Cooling Control
Learning models that minimize the energy use by computer room air conditioners (CRAC) have been in development for nearly a decade and many now work very well. These models consider much more data in their decisions than humans possibly could, and they can produce solutions that humans might never have considered.
Cross Domain Operation Optimization
New software solutions leverage digital twin technology to create a single hub of real-time, trusted asset information. It connects disparate data feeds from the power, cooling and IT domains of the data center to deliver transformational performance improvements in efficiency, reliability or sustainability.
Data Center Design And Construction
AI can be used in the early stages of a construction project to help with design decisions and project planning by analyzing data from previous projects. It can also provide insights into best tendering and purchasing practices, material selection, delivery and installation.
Maintenance
AI can be used to dynamically improve maintenance processes by analyzing data from IoT sensors and equipment data. In a data center, AI monitors the wear of perishable equipment like UPS batteries, fans, and motors, and it can predict when maintenance is needed, as well as failures and end of life.
Robotics
According to Gartner, advanced robots with AI and ML capabilities will be deployed in half of all data centers by 2025, resulting in a 30% increase in operational efficiency. Robots can navigate massive data centers to provide visibility on environmental issues (smoke, leaks, fires, falling over IT racks, high or low temperatures, toxic gas, etc.) that a single maintenance person or static camera cannot replicate. Robots can also perform repetitive maintenance tasks like server replacements in the future.
The Importance Of AI In Data Centers
AI is certainly ingraining itself in almost every aspect of our lives—working, learning, shopping, traveling, playing, and more—and it is all dependent on data centers. The dichotomy is that AI is being used within the data centers that provide enhancements to our everyday lives.
While I predicted AI would be used in the data center, I underestimated the importance of AI in making those data centers secure, efficient, resilient and sustainable.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Read the full article here
