Artificial intelligence (AI) is being unleashed on business processes, data analytics, and a host of other enterprise functions, but its role in data center automation stands to change not just the data center itself but all its infrastructure — physical and virtual, to the edge and beyond.
As with most everything AI touches, the data center will become leaner, less costly to operate, and achieve higher performance metrics as the transition unfolds, and much of what is done by human operators will be automated — just like what was happening in the pre-AI era. Still, experts predict that the shortage of qualified data center operators will continue and may even get worse.
AI of all trades
Clearly, there are many ways in which AI can be used to automate data center management. Rohan Sheth, of colocation provider Yotta, highlights some of the low-hanging fruit, such as load management, power and resource consumption, and security. Still, he says, few organizations are leveraging AI to its full potential, due to a combination of mistrust, lack of skill sets, and the potential risk for disruption.
This won’t last forever, of course. Citing research from Gartner, Sheth notes that by mid-decade the transition from big data to small and wide data will be in full swing, and this will put pressure on data centers to become more flexible and less costly to operate. Rather than simply crunching through extremely large data sets, AI will use lower volumes of more precise data that will be culled from multiple sources and will reside in multiple formats.
While the trend toward intelligent automation was on the upswing before the pandemic, the sudden lockdowns and work-from-home culture that arose in response have kicked the movement into high gear, says tech entrepreneur Don Basile. Over the 2020/2021 time period, 83% of organizations increased their budgets for AI and machine learning (ML), and this has already hastened processing speeds by 30% and has brought data bottlenecks to record-low levels. In part, this is why research houses like Mordor Intelligence are predicting the market for data center automation as a whole to more than double to nearly $20 billion by 2026.
IT job growth
Automation is a scary word for any workforce, and the addition of AI to data center management is causing no little concern in the IT department. But according to a recent report from the Uptime Institute, the data center will likely face a continued worker shortage even as AI becomes more prevalent in the management stack. In fact, demand for human operators is set to increase from today’s 2 million to 2.3 million by 2025, even as upwards of 90% of IT executives expect AI to take on many of the rote, repetitive tasks of data and infrastructure management.
Among the top challenges that IT will face: managing the growth of data loads while maintaining consistency and controlling costs, something that AI can assist with but cannot handle on its own.
And while AI is poised to infiltrate many aspects of the data center, the fact remains that much of the physical infrastructure is still not ready for the change. A recent Gartner report noted that much of the mechanical infrastructure is still lacking the kind of sensor-driven monitoring capabilities that are essential to intelligent control. This means much of the data center infrastructure management (DCIM) stack will likely remain on a manual footing for a few more years at least, even as data loads and resource utilization will climb due to the advent of 5G and the IoT-connected edge.
All of this points to a pretty good outlook both for AI in the data center and the IT workforce. Continued job shortages will likely drive salaries up, particularly for those who have augmented their resumes with AI skillsets, while demand for increasingly sophisticated intelligent platforms proceeds unabated over the next decade or longer.
At the same time, data users in the workplace and consumer markets should see dramatic improvements in speed, accuracy, and overall simplicity once AI takes much of the complexity out of the user interface. And the modern data center itself will be able to do more while consuming less, which is a benefit to all.