As the demand for digital infrastructure continues to increase, new-build data centres and a large estate of legacy IT facilities will be vital in underpinning the ongoing demand for digital infrastructure. Andy Connor, EMEA Channel Director, Subzero Engineering, discusses how we can best meet today’s sustainability challenges, alongside the imperative of reducing energy costs.
In an ideal world, it would be nice to imagine that any data centre over a certain age would be quietly taken offline and retired – many being inefficient and ill-equipped to deal with today’s increasingly high-density IT demands. However, in the same way that the switch to electric vehicles allows for a long transition period away from fossil fuel-powered transport, to cushion the impact and the cost of such change, the reality of the digital infrastructure sector is that legacy data centres will be required for the foreseeable future.
All such facilities are capable of being replaced in a highly ambitious five-to-ten-year period, but the investment required would be prohibitive. Closing older data centres without replacing them would mean a massive drop in digital infrastructure capacity, at a time when the demand has never been so high.
Establishing the exact global data centre population is not easy. In terms of what might be described as the number of data centres owned and/or operated by professional data centre providers, the US International Trade Commission estimates, as of May 2021, a total of some 8,000 data centres globally.
However, when considering the hundreds of thousands of enterprise-owned data centres globally – everything from a cupboard with a few servers, right up to large-scale buildings – then the extent of the legacy problem is truly revealed. Let’s not forget that for every greenfield data centre, there’s at least one older facility upon which one may rely.
Why does all of this matter?
Today, data centre PUEs have come down from an average of 2.5 in 2007 to 1.55 in 2022, but this current figure is still a long way from the ambition of 1.0. If the average is 1.55, that means that there’s a substantial number of data centres that are not even close to that figure as of now, the impact of which comes at a significant financial and environmental cost.
Energy costs for the industry have risen massively since January 2021. While the rate of the increase varies from region to region, many countries have witnessed a more than doubling of the cost of electricity. The silver lining to this energy crisis might be the accelerated development of renewable resources for longer-term cost and environmental benefits – the reality is that the price of power has become unsustainable.
In the data centre sector, where much of the cost of running a facility is the power bill, the impact has been significant. Smarter operators will pass power costs directly on to customers, which on the one hand provides a haven from volatile energy costs, but on the other risks customers moving their business to cheaper or more energy-efficient colos.
As for enterprise data centre facilities, the rise in energy prices across Europe is equally, if not more, impactful than it is for the colo and cloud operators. After all, enterprises are unlikely to be able to negotiate the long-term price agreements that have, to some extent, cushioned the energy crisis blow for large-scale operators.
For legacy data centre operators, the energy crisis is even worse. They are not only paying hugely increased power bills but are paying an additional premium through the inefficient use of power. For example, they waste more than 60% of cooling energy, meaning higher power consumption and a much higher energy bill than is necessary. Thanks to inefficient data centre design and a lack of truly sealed, air-tight aisles of racks and cabinets, the cooling solution deployed in many data centres is having to work for much longer and harder than it would need to do in an optimised environment.
Containing the problem
The good news is that legacy data centres have a roadmap which can take them a long way on the journey towards minimising their carbon footprint and improving energy efficiency. The first step is to analyse and understand what’s going on in the data centre. Without such insight, any attempted improvements will be little more than educated guesses with no real idea of their likely impact.
CFD analysis provides a comprehensive map of how a data centre operates in terms of how effective the existing cooling technology is. It highlights key inefficiencies, such as hot spots, providing an overall picture of where hot air leaks in and/or out of the IT racks, cabinets and aisles, demonstrating the inefficiency of the existing cooling solution. Armed with this picture of current legacy data centre performance, it is then possible to implement a programme of relatively simple but effective measures to bring about a significant improvement.
Most notably, the implementation of a comprehensive containment strategy brings with it significant data centre operational efficiencies. These efficiencies translate into significant financial and environmental benefits, saving as much as 30% of the energy bill and offering a reduced carbon impact. The emphasis is on ‘comprehensive’. For example, there is no point in building a containment system around the IT cabinets, if the power cable ingress ducts are not fully sealed as well. Similarly, if a cabinet is not fully populated with hardware, the containment solution will not be effective unless the gaps are closed with blanking panels.
When properly designed and installed, containment systems can deliver huge operational and energy benefits to the data centre. Set alongside other measures, such as an optimised data centre configuration, increased use of free cooling where possible and/or increased operating temperatures – as allowed for in the ASHRAE standards – containment systems can improve legacy data centre performance dramatically.
Finally, comes the ongoing monitoring and maintenance requirement. Data centres are dynamic environments. The contents of racks and cabinets, hence the power supply required, will change over time. People will go in and out of the aisles, carrying out various tasks and inadvertently or otherwise, impact the space. The performance of the IT hardware may also change over time, likely generating more heat.
The climate outside the data centre will impact the environment inside. In short, there are many ongoing variables in terms of data centre performance, and, at the most basic level, the IT load could vary dramatically on a daily basis.
Making modernisation work
Measure, contain and monitor is the mantra for an effective legacy data centre modernisation programme. Whatever the size of your data centre, understanding its current operational performance gives you a baseline from which to plan any improvements. Any such improvements will need at their core a commitment to a comprehensive containment solution. Ongoing monitoring and maintenance will ensure that the newly upgraded and optimised legacy data centre continues to run as energy efficiently as possible.
Optimising performance from an energy efficiency standpoint brings with it much-needed financial savings at a time of energy price volatility. Further, by using less energy, not only does a data centre cost less to run but also enhances its sustainability credentials. Modernised and optimised legacy data centres can continue to support the ongoing digital revolution and be seen as part of a sustainable, green future and will no longer be seen as the industry’s environmental Achilles heel.
