The Stack Archive

Is it hot in here… or is it just your data centre?

Thu 18 Aug 2016

Data centre cooling

jeff-klausJeff Klaus, GM of Intel DCM looks at how to leverage analytics in every part of your management…

As companies shift their focus on enterprise-wide green initiatives, IT and facility managers are looking for long-term solutions to decrease their global footprint. With 10% of the world’s power consumed by data centres (that’s 91 billion kilowatts of energy) and, of that 10% half is used by servers alone, even when they are idle data centres stand as the perfect entry point to tackling green initiatives.

With green initiatives, many data centres are shifting toward implementing new tools and solutions to boost efficiency in a big way, to do their part to reduce the massive amounts of energy they use. What many don’t realise – which is evident in that more than 40% of data centre managers use manual processes for forecasting and planning – is that analytics plays a role in every step of the data centre management process and can result in saving time, energy and money. From start to finish, managers can use analytics to give themselves the upper hand in building and executing an efficient cooling plan.

Before issues occur

Screen Shot 2016-08-18 at 11.24.14Unfortunately, it’s easier to know what to prevent after it’s already happened. But with predictive analytics, detection of unseen issues before they appear can be highly beneficial. Using real-time telemetry data, administrators can proactively control their data centre based on information provided. A constant read of temperatures, for example, can help managers adjust accordingly to prevent overheating when things get too hot, or to save money and energy by raising temperatures when servers are not at risk of overheating.

During the action

When problems strike – and generally they come without warning – they can devastate an entire data centre. During thermal snafus, managers need to act quickly and accurately. With 57% of data centre operators saying they’ve experienced thermal-related challenges within their data centre, facility managers need to have immediate and actionable data to resolve issues. For example, as heat rises, the servers at the top of a rack may be at higher risk of over-heating. Pin-pointing these issues immediately can impact the amount of time spent working to find and fix any issues.

After the storm

Hindsight is a wonderful thing, but the most important aspect is learning from the data. Following an issue in the data centre, data can provide insights into what key adjustments need to be made to better prevent and prepare for similar issues. IT managers and data centre operators can learn from real situations and incorporate those insights into long-term cooling initiatives, thus optimising efficiency and preventing major malfunctions.

After all is said and done, more experience is gained when learning how issues come about. With real-time data, managers can review and implement what they learn, based on the needs of their systems. Real-time analytics are a viable solution to provide a better understanding of how to achieve improved Power Usage Effectiveness (PUE), and reduce costs associated with cooling and power while mitigating an organisation’s carbon footprint.



Data Centre feature green
Send us a correction about this article Send us a news tip