Preparing data centres for reoccurring heatwaves

It seems like every UK summer is becoming the hottest one yet on record, with scientists warning 2023 will be no different as the El Niño climate phenomenon is set to return. For the data centre industry, another season of temperatures above 40°C challenges us to keep the data halls cool, ensure uninterrupted connectivity, and ultimately, minimise the risk of downtime for organisations. All whilst considering clients’ increasing sustainability requirements. With heatwaves becoming the norm, what should data centre providers be doing to best serve their customers as part of their heatwave mitigation plans?

Taking the lessons from the 2022 heatwave

2022 truly put IT infrastructures to the test, prompting data centres to review their strategies in a bid to prevent disruption in the future. In fact, a recent Climate Crunch research report commissioned by Telehouse revealed that nearly two in five (39%) organisations experienced an average of 10 hours of downtime as a result of the 2022 heatwave, with a further 21% citing 11-16 hours of downtime. Many valuable lessons can be learnt from the events of last year, all of which can help improve processes and procedures for the next hot years ahead. Part of the preparation to ensure resilience throughout very hot summers, and equally freezing cold winters, includes putting the design of data centre buildings under the microscope.

The design of older data centre facilities proves tricky when it comes to minimising service outages during extreme weather. This is because they were built with much lower temperatures in mind, up to 35-38°C, meaning even the smallest of rises outdoors can make a big difference inside. It is therefore a challenge to retrospectively fit and prepare them for extreme heat. It’s not just high temperatures that need to be considered though. Events like Sudden Stratospheric Warmings (SSWs), which can cause cold conditions in winter, are also being affected by climate change. The industry needs to be prepared for these colder temperatures in winter too, all the while meeting ESG and sustainability targets.

It’s not all doom and gloom, however. There are several solutions data centre operators can implement to prepare older facilities for the conditions facing us in the years ahead. These include reusing waste heat, utilising high temperature liquid cooling systems and low power servers, and leveraging hot and cold aisle containment. Many of today’s systems also have a dual purpose making the building more energy efficient. For example, liquid used for immersion cooling as it is a much more efficient conductor of heat than air.

It’s not what you do in the summer, but all year round

With the right infrastructure in place, processes and procedures are the next consideration. Before the start of the summer and winter seasons, it is of course important to double-check whether cooling units are in peak condition, and spare parts available should the need for repairs arise. In winter, some of the cooling equipment can be turned down to ensure energy efficiency. However, more focus should be placed towards conducting year-round maintenance, servicing, and testing with changing seasons in mind.

This will help prepare the equipment for the extreme weather stresses, including chillers, glycol systems, and humidifiers. Additionally, having a schedule of regular servicing and preventative maintenance work can help to avoid major works being conducted during extreme weather events, as it’s usually during these events that supply issues can occur. A checklist of all available resources and historical faults increases the chances of pre-empting potential failures, giving engineers enough time to diagnose to prevent overheating and outages. The key lies in the right planning and preparation.

Embracing a proactive approach

By now, data centre operators know to expect the unexpected and be prepared for every eventuality. When extreme temperatures strike, a well-rehearsed plan based on a proactive approach helps to swiftly minimise downtime risk and resolve any issues.

With all non-essential maintenance postponed during heatwaves, data centre providers can better and more strategically allocate their emergency reaction teams of engineers. The older mentality of ‘business as usual unless something breaks’ isn’t the right one, as every minute of downtime can have damaging consequences for data centre customers. Teams should be able to jump on issues as soon as they occur, hence where they’re stationed does matter. For example, a reactionary engineer should be ready on the roof to fix the cooling systems, if needed. Whilst allocating roles to on-site personnel ahead of the heatwave or arctic blast, data centre providers need to implement extra health and safety measures to help everyone stay productive and safe, with the provision of food and water, appropriate clothing, and of course comprehensive training.

Many colocation data centres have adequate resources available to them to ensure business continuity, even when the wider industry battles skills shortages. For some pieces of equipment, there may only be a handful of engineers deployed by a manufacturer in a certain geographic location. However, more established data centre providers have the size, scale and relationships to ensure priority assistance from these providers in an emergency.

Prevention equals readiness

The importance of prevention and all-year-round servicing can’t be stressed enough. As the entire planet is heating up, the data centre industry needs to take more proactive steps to ensure the right strategy, equipment, and people, are put in place to limit the negative effects of global warming on IT architectures. No two summers are the same, so a high level of preparedness is crucial to prevent issues from arising.