Sponsored: A framework to achieve net zero, environmentally sustainable data centres

Growing concern about the effects of climate change has placed much pressure on the data centre and digital infrastructure sectors. Today data centres are predicted to account for around 1%-2% of global electricity consumption and emit the same level of carbon as the airline sector.

Unsurprisingly, given their importance to the digital economy, the industry is coming under closer scrutiny, not only because of the growing demands it places on power generation and energy consumption, but because of its influence on other areas that affect environmental sustainability. Those which include the appropriation of land and water, the use of renewable energy, and the disposal of hazardous waste material.

Forces influencing the move toward sustainable data centres include a combination of government regulations, the demands of customers and investors, as well as a growing understanding that by leading on sustainability it is interlinked by our desire to grow. For example, in 2021, Schneider Electric and 451 Research surveyed over 800 global colocation organisations and found that 97% of providers’ customers were asking for contractual sustainability commitments. However, while sustainability within the sector is growing in importance, practices are lagging and recent research from the Uptime Institute found most organisations, still, are not closely tracking their environmental footprint.

In all cases, there is a pressing need to measure the effects of one’s efforts, not only because of the familiar dictum that “what gets measured gets managed” but also because of the need to prove to customers, investors, rivals, and regulatory bodies that one is making verifiable, sustainable progress.

The way sustainability is measured also varies from operator to operator, so clearly, there is a need for more reliable, standardised, and transparent metrics to allow data centre operators to benchmark their achievements in as clear a way as possible. All metrics should be meaningful, comprehensive, and verifiable so that the publication of calculated figures will clearly indicate how well a company is pursuing a policy with genuine environmental sustainability benefits, how well the policy is understood, and how much trust can be placed in the veracity of the stated figures.

Energy efficiency

From the first days of cloud computing, energy efficiency was a key concern for data centre operators. As the amount of electrical power needed to operate the IT equipment itself was often matched by that needed to cool it, the cost of running data centres and their effects on overall demand on the public power grid encouraged efforts to streamline the cooling effort for maximum efficiency.

As early as 2007, the Green Grid proposed the metric of power usage effectiveness (PUE), now an ISO-administered standard, as the ratio of overall power consumption of a facility to the power consumed by its IT equipment alone. PUE is now perhaps the most familiar measure of a data centre’s overall electrical efficiency and is widely quoted in benchmarking literature, including that of regulatory bodies concerned with planning and approval for new facilities. According to the Uptime Institute, IT and Power consumption (82%) and PUE (70%) are the top sustainability metrics tracked across the industry.

However, power efficiency alone is not sufficient for the successful implementation of a well-rounded sustainability strategy. On the contrary, a truly holistic effort should consider a multitude of categories and specific metrics to reduce environmental impact. To support the efforts of operators, and advance the industry, Schneider Electric has created a first-of-its-kind ‘Environmental Sustainability Metric Framework’ that empowers the sector to take control of its sustainability goals.

It includes 23 key metrics for operators who are in the Beginning, Advanced and Leading stages of their sustainability journey, helping the industry to standardise the way it measures and reports its environmental impact by proposing five key categories, which include Energy use, GHG emissions, Water, Waste, Land, and Biodiversity.

Taking account of all of these will provide a more rounded picture of how a data centre affects its surrounding environment, and how its owners or operators can monitor the success of their own efforts to drive sustainability. In this article, we’ll explore them further.

Key metrics

Energy is the first category to monitor and measure. As the single most expensive operating cost associated with a data centre, and one that is subject to major price fluctuations due to the geopolitical nature of fossil fuels and renewable energy production, maximising energy efficiency makes both commercial and long-term environmental sense. Reporting energy consumption, energy efficiency and the share of energy consumption that comes from renewable sources is vital for data centre operators keen to reduce their environmental impact.

Second, are greenhouse gas (GHG) emissions. Carbon emissions resulting from the generation of gases such as CO₂ (carbon dioxide), CH₄ (methane), PFCs (perfluorinated chemicals) and HFCs (hydrofluorocarbons) are a major contributor to climate change and efforts to minimise such gases are ongoing across all areas of business. SF₆, for example, is a greenhouse gas over 23,000 times stronger than CO₂ and is found in most existing medium-voltage switchgear, so operators such as Schneider Electric have developed SF₆-free technologies to directly address this.

Water use is another are key to address, and in the United States alone data centres consumed approximately 174 billion gallons of water in 2020. More locally, a 15-megawatt data centre can consume up to 360,000 gallons of water a day so it’s imperative we address this.

Cooling towers and other evaporative cooling techniques are popular methods of heat rejection because of their high efficiency and large cooling capacity. However, the evaporation needed requires the consumption of significant amounts of water. Typically, a 1-megawatt data centre with traditional cooling methods can use about 25m litres of water a year. Also, water is often used in the generation of electricity too. Using reclaimed or recycled water, rather than treated drinkable water, where possible is an important sustainability issue for data centre operators and metrics are needed to guide best practice.

Inevitably, data centres generate waste, both during their construction and operation, which often includes hazardous materials that must be disposed of properly. Circular economy design methodologies, Green Premium technologies and better processes can support improvements in this area. So to can recycling of end-of-life products including batteries from the uninterruptible power supply (UPS) systems. 

Finally, the potentially harmful effects of data centre construction on land use and biodiversity must be kept to a minimum. This is especially true when extra demands on real estate are made, not just by the facility itself but by the associated renewable energy infrastructure such as wind turbines and solar panels.

Choosing appropriate metrics

When selecting metrics for each of the above five categories, it is essential that each one should be relevant and important to data centres, reflecting the impact on the environment directly and indirectly. They should also be easy to implement, at both the data-gathering and calculation stages, and result in benchmarks that are easily communicable, both inside organisations and across the wider industry. Finally, they should lead to actionable outcomes that can drive significant improvements in sustainability, and, where possible, be applicable across all geographies.

By following these guidelines, Schneider Electric has created a framework detailing 23 key metrics to improve sustainability within all five essential categories.

Energy-based metrics

Dealing with energy issues requires data centre operators to first measure the total energy consumption of their facilities, calculate its PUE and measure the amount of energy that comes from renewable sources. Renewables could be located on site, bought from energy companies by purchasing renewable energy credits or through longer-term power purchase agreements (PPAs).

From those measurements a renewable energy factor (REF) metric can be calculated as the ratio of renewable energy to total energy consumed at a site. An REF of 1 indicates that all the data centres power is renewable. Another key metric is energy reuse factor (ERF), the calculation of which is defined under the standard ISO/IEC 30134-6. This is the ratio of energy reused or exported, to the total energy consumption and indicates how well secondary use is made of the facilities energy. A typical example would be how the heat extracted as part of the IT cooling efforts is being used to heat nearby buildings.

Together these metrics can encourage operators to improve their overall energy efficiency, increase use of renewable sources and promote circular economy initiatives such as heat re-use.

GHG metrics

The control of carbon emissions is of global geo-political importance and as such, there are numerous internationally recognised protocols, codified in ISO standards where tracking data centre emissions and those throughout their supply chain including external electricity generation and transportation, is essential.

Many of these are complex calculations encompassing emissions from multiple sources, but they form the basis for the calculation of other metrics such as carbon intensity, which is the ratio of carbon emissions to the total energy consumed by a data centre. Another is carbon usage effectiveness (CUE) which is the ratio of annual CO₂ emissions and the energy demand from IT equipment.

CUE is related to the IT load and allows comparisons of carbon emissions across data centres and other industries. It can be used in the site selection, planning and design phase as well as during operations to measure the effectiveness of continuous improvement programs.

Carbon offsetting and carbon credits, provide a means of encouraging businesses which have more scope to reduce carbon operations to do so, by paying them and deducting the amount of carbon saved from one’s own carbon emissions. These provide economic incentives for reducing emissions globally.

Hour by hour supply and consumption matching will measure the extent to which renewable energy generation matches the energy consumption by an operator. It provides a high level of transparency into how renewable energy production matches consumption in real-time.

Water usage metrics

Key metrics include total site water usage, covering all water consumed – fresh and reclaimed – within a facilities operation. Total source energy water usage measures the water used to produce the energy consumed by a data centre and can be used to optimise the water usage related to energy consumption.

For example, water used by an evaporative cooling system will add to the total water usage but will reduce energy consumption in the cooling effort. This saves water usage at the power plant, and provides a holistic view, enabling better management of all water associated with the plant operations.

Water usage effectiveness (WUE) is the ratio of the date centre water consumption to the sum of energy consumed by IT equipment. Like PUE and CUE, it allows comparisons to be made across different data centres.

Waste metrics

Here key metrics include the total weight of material waste generated at a data centre, from construction right through to operation; the weight of waste sent to landfill sites; the weight of waste diverted from landfills through circular economy efforts including re-use, manufacturing, and recycling; and waste diversion rate, which is the weight of waste recycled, divided by the weight of total waste generated.

This last metric creates a ratio that can be compared across data centres, so can be used to benchmark continuous improvements in waste reduction.

Land and biodiversity

Although there is a general appreciation that the development of land for data centres should not adversely impact biodiversity, including animal habitats, plant life and even micro-organisms, metrics to compare efforts are in their infancy and not yet standardised.

Application of metrics

Once an organisation commits to gathering and processing sustainability metrics, they can be applied across a range of functions to deliver genuine improvements in sustainability. The most obvious is target setting: aiming to improve the performance of a data centre or the overall organisation, whether by aiming to achieve a particularly ambitious PUE, CUE or WUE target, or aiming to reduce the amount of waste generated over a set period of time.

Metrics must also enable businesses to report accurately and communicate their progress towards achieving sustainability in a transparent and measurable way. It should also offer the possibility of certifying their efforts against accepted standards.

With the demands placed on the sector accelerating, and the need for resilience becoming greater, data centre operators must consider how their business choices impact the environment and prioritise standardised, sustainability practices within their organisations.

Leading the way to a more sustainable future

At Schneider Electric, we are happy to introduce our latest White Paper – WP #67 Guide to Environmental Sustainability Metrics for Data Centres, which I believe will lead the charge for more sustainable data centres. Data centre service providers and companies that operate data centres can benefit by using this guide on their data centre sustainability journey.  We go into great detail on all of the categories and metrics in WP #67. If you are interested in innovation, sustainability, and data centres, we encourage you to read it here.