Interestingly, more respondents (55 percent) said data center expansion requires the majority of their efforts compared to migrating to external clouds (42 percent). Jeff Klaus, GM of global data center management solutions at Intel, admits these numbers are a little surprising.
“The wave of public cloud migration has been occurring over the past couple of years, and maybe we are starting to see the pendulum swing back a little bit toward private cloud,” he said. “That’s the only way I can explain it.”
But, he added, while we typically think of data center expansion in terms of adding server nodes, IT managers could also include other elements like upgrading to 10 Gigabit Ethernet, for example. “It could be some upgrades to the system that might not necessarily mean you are expanding private environments but still take a lot of IT’s time,” Klaus said. “Maybe IT modernization might be a better way to explain it.”
Morar Consulting carried out the study, surveying 311 IT directors and managers, IT system administrators, and applications architects in the U.S. and U.K. Its findings show the need for more granular insight into data center power consumption. And it concludes that IT teams are looking to new technologies and methods like artificial intelligence (AI) and DevOps to meet their data center demands.
Eighty-seven percent of respondents said their organizations had experienced power issues. And 40 percent said improving power delivery efficiency is their top goal for managing data center power consumption. For comparison, 35 percent said improving power usage monitoring and 34 percent said improving the ability to anticipate server workloads were their top goals.
“Improving power deliver and server efficiency continues to be a challenge,” Klaus said. “We’ve been preaching the real-time [monitoring system] benefits for so long, and we want to get to a saturation point. But 40 percent of respondents feel they still have some way to go.”
AI for Data Center Management
And, of course, data center management software by vendors like Dell (that co-commissioned the report) and Schneider Electric and Siemens (both Intel partners) use AI to provide this type of real-time granular facilities management while freeing up IT teams to work on other issues.
The survey found that over two-thirds (66 percent) of teams have AI-related projects in place, while 78 percent are starting to use AI capabilities offered by their data center management tools. Additionally, 88 percent said they have deployed or are interested in developing analytic tools to better formulate solutions, and 69 percent said DevOps has changed the tool chain and metrics that their company uses to measure value.
The study, Klaus says, “reinforces the need for awareness and continued education on the efficiency of [data center] monitoring.”
For example, Intel is working with a couple of large cloud companies in China and a few in the U.S. that have major challenges predicting demand, he said. There’s also “a company I won’t name yet but they are one of the biggest car companies that doesn’t make a car, and they have a very difficult time predicting usage,” Klaus said, adding that AI and predictive analytics can help with these issues. But they will also require more memory to meet demand spikes.