Meta is convinced of one thing: the occupational safety of data center personnel must be constantly studied. Research suspects solutions based on the benefits of Artificial Intelligence.
Within data centers, regular maintenance of equipment is never without risk for personnel. Between emergency situations, exposure to products harmful to health and work accidents, everything is possible. Companies then turn to AI to try to find more sustainable solutions to this kind of problem.
Microsoft, Meta and Google: research into AI systems is underway
Such is the case of Microsoft. He has developed an AI system to detect the sources of problems. This solution is also effective in alerting staff to security risks. Another system in progress will prevent the consequences on data center construction schedules. Microsoft has more than 200 data centers across continents, and that number is growing.
On the other hand, Meta, account around 20 active data centers worldwide. This company is experimenting with how AI will enable better management of data centers during extreme environmental conditions. Meta was able to conceptualize hostile environments. The implemented AI models – which manage energy consumption, cooling and air fluctuation within the servers – will then be tested.
Security is not the only concern of companies. We all know that in the digital world, breakdowns are dreaded as much by their cost as by the embarrassment they cause. IT Uptime Institute lays out the statistics during 2020. A data center suffered at least one huge outage in a year. And yet a breakdown will cost a company around one million euros.
AI would enable a greener and more sustainable digital industry?
AI also promises other advantages. Thanks to AI, it is also possible to make energy savings which mean financial reductions for data centers. In 2018, Google revealed that pre-engineered AI systems by their subsidiary DeepMind had seen a 30% reduction in energy consumption. No explanation was issued by DeepMindhowever, other tech giants continue to explore AI to reap its full benefits.
Microsoft had previously used AI systems to detect anomalies in the use of electricity and water within data centers. This could have been done through equipment telemetry data. AI is also used to solve problems with energy meters, and to optimize the cooling system, network and reduce waste. On the other hand, Microsoft aims to have all its data centers cooled with green energy by 2025.
Meta mobilizes the knowledge of reinforcement learning to reduce the amount of air pumped into data centers to cool them. Meta, for its part, has been boasting since 2020 that all of its data centers already run on renewable energy.