Facility Management sits down with Faraz Ali from Vertiv to discuss why now is the time for FMs to consider greener methods of optimising data centres.
Data centres keep business going in the digital world. They power anything from retail point of sale (POS) systems to the artificial intelligence (AI) and analytics employed by enterprises, which means there’s no room for downtime.
The facility managers of data centres have a critical role to play in maintaining uptime. Handling everything from facility access to energy management, security, emergency preparedness, power and cooling, FMs “cover the lot, and a lot”, says Faraz Ali, the ANZ product manager for integrated rack systems at Vertiv.
The biggest challenge currently facing data centre FMs, according to Ali, is striking the balance between return on investment (ROI) and environmental, social and corporate governance (ESG).
“We’ve all seen the headlines – data centres are significant users of power and water. But the social and government pressure to prioritise reducing carbon emissions is an increasingly loud conversation around the data centre,” he says.
“In Europe, FMs are expected to follow strict guidelines on power usage effectiveness (PUE) – the ratio of total facility energy to IT equipment energy. While Australia is yet to implement such a guideline, we’re on the right path with carbon credits and green star ratings.”
The guidelines will however evolve, says Ali, and now is the time for FMs to be looking at greener methods.
The AI problem
The International Energy Agency estimates that data centre electricity consumption currently represents one to 1.5 percent of worldwide electricity consumption. But recent studies have found that AI uptake will likely increase the energy used by data centres in the coming years.
“From what we can see, AI applications are going to need about four times the power that non-AI applications will use. By nature, this will see power cost and carbon emissions increase,” says Ali.
But with this challenge comes opportunity for innovation, like introducing lithium-ion batteries or seeking more efficient thermal management.
Cooling your data centre
Maintaining the right temperature inside a data centre is one of the most important tasks of the data centre manager. It’s also a key domain for reducing carbon.
“Immersion cooling, for example, does an impressive job at reducing footprint by producing more centralised cooling in a smaller space,” Ali says.
In the era of AI, “re-architecture to the cooling domain” is a path Vertiv is seeing data centre operators and FMs increasingly follow.
“This is because one of AI’s many impacts include seeing some traditional cooling methods reach operational limits,” he explains.
Data centre managers and FMs can generally pursue one of three paths:
- integrating direct-to-chip liquid cooling in air-cooled facilities that currently lack the infrastructure to support it or aren’t ready for mass high-density loads
- redesigning air-cooled facilities with new direct-to-chip and/or immersion infrastructure to support demand for liquid-cooled racks, or
- developing immersion-liquid-cooling-only data centres.
“While the pressure to reduce energy consumption is universal, the market is never uniform. Paths chosen will largely come down to AI readiness,” says Ali.
An AI solution
While AI presents a problem for data centre managers, ironically it holds some of the solutions. As AI becomes more common inside the data centre itself, Ali expects to see this capture real savings industry-wide.
“With the ability to analyse every touch point in the data centre, AI can make fine-tuned adjustments to promote efficiency,” he explains.
“In the cooling domain, for example, by identifying [that] uptime can be achieved running thermal equipment at 40 percent, a data centre would be making a huge dent optimising equipment which previously ran at 80 percent.”
AI can instantly diagnose room for improvement and change the ecosystem for the better. But to see this technology truly take off, the industry will need a “change in mentality”.
“FMs are primed to articulate the cost savings and environmental benefits of a data centre autonomously running at optimum efficiency – their consultation in this space during digital expansion discussions is invaluable,” says Ali.
Best practices for data centres FMs
Ali’s number one tip for FMs is to always have an eye on your data centre. It doesn’t have to be your own eye, it should be through your AI and IT management systems.
“Is something turned off? Would you know if something got too hot? Would you know if your batteries lost power? How would you be alerted? Are you getting the best level of monitoring? These are some of the questions I’m asking FMs every day,” he says.
With this level of insight, FMs are better placed to optimise their data centre footprint.
“Most find the biggest cost is their inefficiency to cool, but the tech is available to help. From putting in containment solutions to applying automatic modulation to chiller units, a little bit of optimisation can go a long way.”
Faraz Ali, ANZ product manager for integrated rack systems at Vertiv.
Is predictive maintenance the future of FM? See what two facilities managers had to say.
