5 Applications for Machine Learning in the Data Center

The COVID-19 pandemic has increased the pressure on IT teams to deliver highly reliable services to work-from-home users. At the same time, remote work has increased IT operational challenges, forcing IT teams to manage and maintain IT infrastructure without on-premises access. Machine learning offers a solution to both conundrums.

Machine learning is a form of artificial intelligence (AI) in which systems improve from experience without human intervention. The systems “learn” by applying past observations, examples, or instructions to new patterns in data.

In the data center context, machine learning can analyze information from IT infrastructure components to optimize operations. Machine learning systems can identify patterns that the IT team might not notice and solve problems faster than humans can. The technology also makes it possible to manage systems more efficiently and address potential problems before they cause downtime. Here are some applications for machine learning in a data center environment. 

1. Reducing the Risk of Downtime 

The Ponemon Institute has calculated the average cost of a data center outage at nearly $9,000 per minute, including IT costs, lost revenue, lost productivity, opportunity costs, customer churn, and reputational damage. Reducing the risk of downtime is a top priority for IT teams. The good news is that the root cause of most outages — UPS failure, human error, and cooling system failure — can all be prevented. Machine learning can help by monitoring systems in real-time and automating many routine tasks.

2. Improving Energy Efficiency

You wouldn’t think that Google’s highly optimized data centers could be any more efficient, but an AI system developed by DeepMind was able to reduce Google’s energy usage by about 30 percent. For most organizations, machine learning can enable even greater energy savings. Some machine learning systems are even able to implement changes automatically, such as rerouting workloads to underutilized, energy-efficient servers.

3. Accelerating Root Cause Analysis

When outages and performance problems occur, IT teams need the ability to determine the root cause quickly. Machine learning systems can identify and solve issues in near real-time to greatly accelerate problem resolution. More importantly, they can move throughout the IT environment, applying lessons learned to prevent problems from propagating.

4. Managing Log Files

The ever-increasing number of systems and devices in the data center are continually creating log files, generating so much data that IT teams cannot sift through it all. Machine learning tools can gather and ingest log data into a central location, and analyze the files to create valuable, easy-to-read reports.

5. Enabling the Autonomous Data Center

Machine learning systems that analyze data and generate alerts can deliver tremendous value. Advanced tools go further by automatically implement changes in real-time. Data centers become more autonomous, increasing IT efficiency, and providing for more effective management.

Conclusion

IDC has predicted that, by 2022, half of all data center assets will utilize embedded AI functionality to operate autonomously. This forecast was made long before the COVID-19 pandemic upended IT operations. As IT teams grapple with the challenges of remote work, machine learning systems can take over a wide range of tasks and provide the insight needed to optimize the data center environment.