In this months CISBE Journal, our head of innovation, Richard Clifford, argues that early collaboration is essential to ensure customers understand and avoid the hidden costs of energy saving measures within data centres.
Data centres are inherently energy intensive, making up as much as half of a company’s energy consumption in some cases. Naturally, as financial, regulatory and CSR pressures have increased, energy efficiency has been pushed to the top of the agenda for data centre design. A myriad of options available in the market, exacerbated by different consultants all championing their own approach or solutions, has led to confusion and in some cases an inherent lack of understanding for the end user. In the race to specify efficient data centre estates, options and outcomes are not being fully explored, which can lead to increased costs and compromise overall CO2 reductions in the long term.
Part of the problem is that the industry’s go to metrics can be misleading or easily manipulated to appear more attractive. Specifiers often focus on metrics because at a base level they provide an easy reference point that can be used to evaluate the efficiency of different options. Yet, measures that bring the best results on paper may not offer the most efficient or cost-effective option and can, in extreme cases, lead to long term operational faults.
Most of these measures also work on an assumption that the facility is operating with a full IT load, which rarely happens in most cases. Energy efficiency naturally decreases at lower IT loads and suggested levels can take longer to come into play, if ever. Solely investing in a design that offers a good efficiency metric can leave a business open to higher operational costs than expected later down the line.
For example, PUE (power usage effectiveness) – a ratio of how efficiently energy is used by a data centre’s computing equipment in contrast to cooling and other overheads - can be improved by raising rack temperatures, which has very little effect on savings, as any energy saved by turning down cooling systems is shifted to server fans.
Metrics like PUE don’t paint a full picture on efficiency and, as consultants, we should be working with all our stakeholders to ensure they understand the true implications of how different solutions and routes will affect their facilities in the long term.
Many of the energy efficiency measures available are key examples of this and their long-term implications need to be thoroughly understood before they are committed to. Without proper consideration, some can create hidden costs elsewhere and cancel out any potential saving, or, in the worst-case scenario, have a negative impact on the facility’s resilience, increasing the risk of downtime.
Fresh air cooling, often considered one of the most eco-friendly cooling methods, is a textbook example. On paper it has almost no carbon footprint and can be particularly cost-effective for facilities located in colder climates. Such alluring benefits are causing many customers to specify these systems without considering suitability or being aware that operating these systems long-term can be more complicated than first anticipated. Fresh air cooling is highly dependent on location – it is easily contaminated by pollution or seawater which can damage hardware and bring high replacement costs. Preventing this damage means investing in additional cleaning and maintenance which brings more overheads and ultimately, a larger CO2 footprint if replacement parts are needed. Likewise, fresh air needs more complex control systems such as fire detection and suppression which all add further capital and operational costs.
Another cooling example impacting efficiency are the new changes to F-Gas legislation, which have brought forward price rises related to the management of stalwart gases such as R404A and R410A. This could have a large impact on pumped refrigerant DX systems, and lead to significant maintenance issues despite their lower capital costs.
Collaboration is key
It may be an attractive option to choose methods that appear to represent a lower upfront investment but without full consideration of the operational lifecycle, any financial or carbon savings may be eliminated in the long term. Many end-clients are stung because they don’t take a wide enough view of their data centre’s management, agreeing to efficiency measures without consulting the teams responsible for the day-to-day running of the facility.
Early consultancy and collaboration at the design stages is vital to prevent this from occurring. We should be ensuring that all stakeholders, including FM and operational teams, are involved in the design process to ensure the operation and maintenance of the facility is considered from the outset. This allows any potential pitfalls to be flagged before any decision are set in stone. It also means all teams can work together to argue for models that may cost more initially, but offer greater long-term efficiency and business flexibility.
Efficiency will always be a concern for the industry, as it should be. But this needs to go beyond just meeting metrics. By considering a facility’s lifecycle rather than just initial costs, environments can be created that save money and energy throughout the course of their use. As consultants we need to be driving this shift in focus to ensure data centres are built with long-term operation in mind.
This article was first published in the April edition of the CIBSE Magazine