Data Centres: A Case Study in Job Creation Aided by Automation
by Clifford Federspiel, President & CTO at Vigilent Corporation
Technology, specifically automation, was called out as a job killer by MIT Sloan School of Management professors Erik Brynjolfsson and Andrew McAfee in the MIT Technology Review.
Their research showed how middle class jobs can be erased by automation. Any casual observer can see how this affected the vote in key Electoral states in America’s industrial heartland during the recent US Presidential race.
In contrast, the Wall Street Journal recently profiled a Deloitte study stating that automation can drive increases in jobs, and can alter the quality of jobs – mostly in a good way.
So, where’s the truth? Does automation help or hurt workers? As is often the case, it depends. Slow-growth industries use automation to reduce costs in the face of stiff price competition, displacing workers in the process.
Meanwhile, high-growth industries mostly use automation to deal with the challenges that high growth creates, making workers more valuable and necessary.
In many cases, automation is assistive. Assistive automation increases productivity without eliminating jobs. The increased productivity enables job creation in other areas.
Fortunately for us, the data center industry is in a hyper-growth phase that makes it an oasis of opportunity and job creation, enabled by advances in assistive automation. In fact, job growth aided by assistive automation is the only way that data centers can possibly keep up with the growth and complexity they face.
And yet, despite the exponentially increasing demand for data and resulting impact on data center operations, parts of the industry have been slow to embrace change, for example:
- Capacity investment decisions often rely on tribal knowledge and best practices that are only “best” in the absence of information that is available today from actual operating data and analytics. For example, a decision to add IT load to a facility should be based on the actual, measured capacity of the existing cooling infrastructure. Doing so allows businesses to grow faster when sufficient cooling capacity already exists, while avoiding the risk of overprovisioning a data hall with too much IT load.
- Maintenance is often performed on a fixed schedule. But maintenance incurs both cost and risk. Performing maintenance when analytics indicate that maintenance is actually warranted helps reduce cost and reduce risk.
Automate or Be Left Behind
Traditional modes of operations management as described above are not sustainable in an industry facing explosive growth. Data centers now operate with greater variability in server density and migration of IT load between facilities, creating complex and interdependent systems that long ago surpassed the ability of humans to manage manually.
To deal with these complexities, hyperscale leaders like Google and Facebook have already fully embraced the use of automation technologies in their data center operations.
Facebook has been amassing robotics engineers with data center experience as it simultaneously expands its massive build-out of server farms, and is collaborating with others with the Open Compute Project. Google, which joined OCP this year, has used their DeepMind machine learning to reduce cooling by 40% in their already-efficient data centers.
It’s my view that data centers in other sectors — colocation, telecom, enterprise and government — benefit from automation even more than hyperscale operators.
Why? Because hyperscale companies have an abundance of technical talent that they can apply to optimization, while other sectors face a shortage of data center engineers. Automation can fill the gap, make jobs more productive and enjoyable, simultaneously opening up new opportunities for growth and employment.
Better Jobs, and More
Running mission-critical infrastructure is complicated and difficult, and continues to get more challenging as the number and size of these facilities grows rapidly. Large data centers can have thousands of racks supported by hundreds of cooling units. Large network operators have thousands of “edge” facilities that typically run “lights-out.”
These facilities are as critical to the health and welfare of our economy as our electrical grid and our highway system. The problem is that there aren’t enough operators and engineers to keep up with the growth rate and the round-the-clock service requirements, which is stressing the workforce and limiting growth.
Assistive automation can help solve these challenges. Controls assisted by machine learning can make minute-by-minute operating decisions, and at the same time provide operators with data-driven insights so that they can more reliably schedule and perform maintenance, and deal with expansion and change management.
The data that’s collected by sensors and software can be used to create predictive and prescriptive analytics, which can be used to inform decisions about capacity, reliability and efficiency.
New hires – data analysts – are needed to take full advantage of these analytics. Automation can handle minute-to-minute tasks that require round-the-clock vigilance, and assistive automation can help operators, planners, engineers and analysts with decisions that only humans can make.
The point is that automation helps data centers run better, makes current employees’ jobs more productive and enjoyable, and provides high-value information that requires incremental hires to utilize the information to further optimize the company’s facilities.
The Future is Bright
As the Deloitte Study states “… the last 200 years demonstrates that when a machine replaces a human, the result, paradoxically, is faster growth and, in time, rising employment.
The work of the future is likely to be varied and have a bigger share of social interaction and empathy, thought, creativity and skills.”
Nowhere will this scenario be more likely to play out than in the data center industry. The sheer scale of growth facing the industry requires automation technologies to deal with the complexity of operations and planning.
Automation technologies need more people, with more skills and in more interesting jobs, to deliver on their promise.