Keeping data centers staffed isn't easy and it isn't getting any easier. For one thing, the staffing needs of data centers are complex, ranging from facilities people to keep the power supply and cooling dependable, to people at the cutting edge of networking technology to keep the data center doing what data centers are supposed to do -- keep data flowing.
Even vendors understand these staffing issues and are quick to advertise that their latest technology, from DCIM solutions to AI infused hyperconverged systems, will help data centers keep staffing needs to a minimum.
Maybe to a minimum, but data center staffing needs aren't going away any time soon, no matter how much machine learning you deploy.
There was a time when staffing data centers was much easier. Data centers were few, and the tech job market was filled with fresh community college and university grads ready to jump on the emerging tech bandwagon. It was also a simpler world. There were no virtual machines, containers, edge devices, internet and the like, meaning data center operators' needs were much less refined and defined than today.
"We, as a industry, are young in that we formed in the last 20 or so years," Lee Kirby recently told Data Center Knowledge. "Those that started in the industry are now starting to retire. Through normal growth we'd be able to handle that, but with an industry that's doubling in size every four or five years, that means a doubling in requirements as well. "
Kirby brings enough data center experience to the table to know more than a bit about staffing them. Since 2016 he's been the president of Uptime Institute, the advisory organization that developed the widely adopted tier standard for data center design, construction, and operational sustainability, and focuses on improving data center performance on an industry-wide level. Before that he spent nearly three years as the organization's CTO.
He also co-founded Salute, an organization that hires and trains veterans for data center jobs.
"There's not one typical data center that people have to support out there," he said. "We've got the large hyperscale data centers that are hundreds of thousands of square feet. We've got the small data centers that are being deployed at the edge. From the core to the edge, you've got all different form factors, all different types of configurations, different requirements, and the complexity of the environment has gone up."
According to Kirby, in today's environment data centers should take a three pronged approach to addressing staffing issues, starting with conducting a staff evaluation to identify current gaps and mitigate future issues.
"Used to be you could get by with hiring good mechanical and electrical engineers, get your OSHA safety requirements satisfied, and tick the box on training because things were a lot simpler then," he explained before recommending that data center organizations perform an assessment to discover and understand the skills they have in place. "Once they know what they have, then they need to match that up against what skillsets they need to have in place."
The second prong is investing in the right training programs. There's much to consider here, Kirby said, because consideration must be given to offering the right training for people who are already on staff as well as for future hires holding positions that might not currently exist in a particular data center. While some training can be done in-house, much will need to come from outside sources such as local community colleges, private training companies, as well as colleges and universities.
"It's gonna come down to budgeting," he said. "One, you need to budget for the training, but two, you need to spend your money wisely. Prioritize training on what will have the biggest impact both short term and long term. That goes back to understanding where your gaps are, because the gaps you identify are going to identify your risks too. You'll base your training initially on the high risk items, then on the long term risks."
Kirby pointed out that some of the necessary training might not always be obvious.
"Many times some of the softer skills are being overlooked," he said. "If you have staff that have been in your data center a while and now they're getting senior positions but they don't have good writing and communication skills, they're going to fall far short when it comes to setting up programs that involve external service providers, where they need to put very clearly in contract terms what it is they need for them to provide."
The third prong of Kirby's approach is probably the most important, especially as far as long term sustainability is concerned: set up and fund pipelines that will drive quality hires. For this stage, he said, data centers will most likely utilize many of the same organizations that they use for training.
"I think the data center industry is a little bit invisible to colleges and even tech schools. It's starting to get more visibility. At Uptime we're working with local colleges to even acquaint their STEM students to the fact that there's an entire industry out there that they should be considering."
Community colleges also offer a great opportunity for getting future hires in the pipeline, because of their mandate to provide vocational training to meet the needs of their local economies.
Incorporating universities and community colleges into the feeder system might also help the industry in another way as well.
"This industry is a male dominated industry," Kirby said. "The more we can do to attract females to the industry, the more we tap into an as yet unrealized resource pool. When you look at the percentage of females in our industry, it's even less than what's in the military.
"We should think about the different resource pools we're tapping into. What we're trying to do at Uptime is work with the students, work with the different resource pools, make them aware of the opportunities, and see if we can attract them into it."
Kirby is confident that if the data center industry can follow this threefold process it can find a way to meet staffing needs in a sustainable manner. The option of data centers robbing talent from other data centers to keep operations running isn't much of a long term solution alternative.
"The revolving door is just a band-aid that's not even covering the wound."