How to lower IT’s digital carbon footprint
The environmental effect of technology is fast becoming a key driver — and drag — on every company’s sustainability journey.
Many industry analysts and IT leaders are turning to technology to help them grapple with carbon accounting and other environmental efforts. But less well-publicized is technology’s high negative environmental cost. Business and IT leaders would do well to understand this complex issue and take action now. Understanding and addressing an organization’s digital carbon footprint is a key component of this approach.
What is a digital carbon footprint?
A digital carbon footprint refers to the greenhouse gas (GHG) emissions that digital technology resources, devices, tools and platforms produce.
The environmental effects of the internet and digital technology tends to be out of sight, out of mind. Nevertheless, those environmental costs are real. A digital carbon footprint includes the following:
- device and equipment usage,
- the emissions produced in their manufacture,
- the energy required to run devices and equipment,
- data transfer and
- the energy consumed by servers and data centers.
These are just some digital carbon footprint sources:
- cloud computing,
- online shopping,
- online games,
- social media,
- mobile phones,
- video streaming,
- Cryptocurrency and
- metaverse platforms.
Digital technologies are responsible for 3.7% of global GHG emissions, according to the report “Lean ICT: Towards digital sobriety,” published in March 2019 by The Shift Project, a Paris-based think tank. Digital transformation — with its reliance on servers, networks and terminals — is driving this trend. The direct energy footprint of information and communications technology (ICT) is increasing 9% annually, with video streaming one of the biggest contributors to these numbers.
Assessing a digital carbon footprint is complex, but even research that puts technology’s GHG contributions at a lower amount finds that technology is a significant contributor to global GHG, so companies and individuals must work to lower those carbon emissions.
The complexity of a digital carbon footprint
Many CIOs and technology teams have added sustainability to their agenda, especially with CEOs and boards prioritizing this critical issue. But before IT can lower its carbon footprint and become more sustainable, leaders first need to understand current emissions and harmful practices. At many organizations, sustainability data is scattered among different departments, often in Excel sheets and unstructured and structured data without any alignment. And knowing how to measure CO2 is even trickier.
For many CIOs, there’s a gap between wanting to be green and being green, according to the 2021 report “Sustainable IT: Why it’s time for a Green Revolution for your organization’s IT” from Capgemini, an IT consulting firm headquartered in Paris. The following report statistics underscore the need and complexity of greening IT:
- In 2019, 53.6 million tons of e-waste was produced, a rise of 21% in five years.
- A full 89% of organizations recycle less than 10% of their IT hardware.
- Fifty percent of companies said they have an organization-wide sustainability strategy, but only 18% have a complete sustainable IT strategy with clear goals and timelines.
- Only 43% of executives said they know what their organization’s IT footprint is.
- Only 6% of surveyed organizations have a highly mature sustainable IT.
- Although 33% of organizations mandate IT carbon footprint reporting, most don’t require their vendors to disclose carbon footprint data.
- Many organizations find measuring the IT carbon footprint to be a complex and difficult task, with 49% citing the lack of access to common measurement standards and ratings and the difficulty of understanding the many layers of IT’s footprint — email, voice, video and all other digital technologies.
Steps to lower your digital carbon footprint
Despite the complexity of a digital carbon footprint, organizations need to begin the steps to lower it. Here are some steps that can help.
1. Begin with the easier tasks
Without a clear picture of an organization’s digital carbon footprint, IT leaders may be understandably confused about how to start an initiative. CIOs and IT teams should begin with obvious action steps, in other words, “low-hanging fruit.”
One of the first things organizations can do is to control device use and try to do more with fewer devices, while maintaining efficiencies, said Benjamin Alleau, global head of the group sustainability portfolio at Capgemini.
Benjamin Alleauglobal head of the group sustainability portfolio, Capgemini
“The first step is about reviewing processes and the way that people are using those devices,” Alleau said.
Much of this comes down to increasing device lifecycles, he said. Many organizations replace devices every three years. Extending that period to four years reduces carbon footprint significantly. However, this step can be challenging: To maintain efficiencies, the equipment must continue to receive vendor support and be compatible with new software upgrades.
“We have to make sure that together, with the suppliers, we can achieve this target,” Alleau said.
2. Understand emissions categories
Every IT leader should understand the fundamentals of emissions to help their company on its sustainability journey. Understanding what scope emissions fall into is a starting point.
The Greenhouse Gas Protocol, a standards body headquartered in Washington, D.C., divides carbon emissions — digital or otherwise — into three categories:
- Scope 1 includes all the direct emissions a company generates such as emissions from company vehicles, processes, greenhouse gas leaks and manufacturing.
- Scope 2 includes indirect emissions from purchased energy, such as the electricity purchased to power offices and devices.
- Scope 3 includes all the indirect emissions associated with a company’s operations produced along the entire supply chain among an organization, its suppliers and its customers. This includes the raw materials and precious metals used to manufacture digital devices.
“Every organization is challenged by Scope 3 emissions because you need the support of all parties in the supply chain,” said Alexandra Knupe, global head of corporate social responsibility at Atos, a global digital transformation consultancy headquartered in Bezons, France, that offers decarbonization services.
While carbon footprint tools provide emissions calculations for specific categories such as IT hardware, more detailed information would be ideal, Knupe said.
“In the future, it will become more interesting to know, in more detail, what is really impacting the carbon footprint,” she said. For example, what’s the digital carbon footprint of a specific PC terminal, housed in a specific environment, and in use for a specific number of hours? That’s one example, Knupe said.
“The better data we collect, the easier it is for us to offer real decarbonization targets,” she said.
Enterprise systems are always at the ready, and that comes with a climate cost. IT leaders must examine the sustainability of their networking practices.
One of the main reasons data exchanges are carbon-intensive is because unlike mobile devices, which have a “sleep” mode, networks are always on, said Asim Hussain, chairperson and executive director of the Green Software Foundation, a London-based nonprofit organization whose mission is to create and promote standards, tooling and best practices for building sustainable software.
“A network switch, even if it’s not being used, needs to be on and it needs to be consuming what would probably be maximum energy, even if no data is flowing through it,” Hussain said.
This is because data, such as video, doesn’t tolerate latency, he said. “When you transmit data, you can’t be waiting for a [switch] to wake up; it needs to go.”
One potential solution being researched is how to apply the principles behind delay-tolerant networking, Hussain said. For now, a more accessible approach is to turn to energy-efficient hardware when it’s time for a device refresh, he said.
3. Optimize data storage and cloud use
The rate at which large data centers consume global electricity is rising by 10% to 30% annually, due to continually increasing workloads, according to the International Energy Agency. Those percentages would be far higher if the measurement included cryptocurrency mining.
Energy consumption required for data storage is a function of how quickly users need to retrieve it, Hussain said. If the answer is “immediately,” the storage system will be less green. A tiered storage system that could address this sustainability challenge might look like this:
- Organizations store data that users don’t need to access immediately, such as old financial records, on a long-term storage medium that might take time to spin up.
- Organizations use a more frontline system to store data requiring instant access.
Creating this layered approach to storage requires companies to have a clear understanding of the data they collect and store, which can be a tall order, Hussain said. The tendency is to hold on to everything and keep it within immediate reach.
But as the economy contracts, organizations will streamline their storage policies, if only to manage costs, Hussain said. The sustainability benefits, while perhaps not the primary driver, will be a positive side effect.
4. Consider the cloud vs. legacy tech issue
The cloud runs on data centers, which have massive negative environmental impacts. Despite cloud providers net-zero pledges and sustainability claims, data centers require electricity, need water for cooling, generate noise, use hardware and generate e-waste. That’s why reducing environmental effects in ways that don’t create negative material is important. Once IT leaders address carbon emissions reductions, moving to the cloud might be part of a company’s sustainability strategy, according to many analysts.
From a business-only, non-sustainability perspective, the primary appeal of cloud computing is that it’s based on a pay-as-you-go model, said Owen Rogers, research director for cloud computing at Uptime Institute LLC, a digital infrastructure standards and certifications body headquartered in New York. When companies are strategic about how they invest in cloud services, they can also reduce digital carbon emissions. Rogers’ research focuses on how to use cloud services to drive business value while achieving sustainability at the same time.
“If you’re using fewer resources, it means there are less virtual machines being spun up in a data center consuming power and [emitting] carbon,” Rogers said. “[And] if you have virtual machines that are highly utilized and every single bit of that virtual machine is driving value, you’re obviously operating more efficiently than if you have a virtual machine that’s not really doing anything.”
In other words, the best practices that apply to cloud spending are also applicable to a reduction in carbon emissions.
This works when applications are scaled up and down — again, depending on use. For example, an e-marketplace may scale up cloud use during the holiday season to accommodate more traffic on its website, and then scale down during the slow period.
Companies might not have been using legacy applications in a scalable and energy-responsive way.
“There are some legacy applications, and some poorly designed applications, and these are very fixed: They always consume the same amount of resources,” Rogers said. “What that means is a lot of those resources are going to be doing nothing some of the time because there’s no demand for them.”
This approach is akin to keeping a server plugged in all day just in case it’s required, he said.
Owen Rogersresearch director for cloud computing, Uptime Institute LLC
“A really good cloud application should … only consume what it needs at the time,” Rogers said.
The first step in decreasing the cloud-related digital carbon footprint is to ensure applications are well-architected, he said.
5. Build greener software
One reason some software might not be as efficient as it could be is because it isn’t designed to work on battery-powered devices, such as mobile phones, Hussain said.
“If you’re building software for something that’s plugged into a wall socket, there’s almost zero pressure to consume less energy,” he said. “In the cloud space, for instance, all those machines are plugged into the grid, so there’s no pressure for cloud engineers to consume less energy because they’re like, ‘we don’t even get measured on that.'”
To create greener software, developers should turn to mobile computing application development best practices — in other words, develop as if the applications must optimize battery use, Hussain said.
“This knowledge already exists, we just need to bring it into a different domain,” he said.
E-waste is another critical issue.
CIOs and IT leaders should demand that vendors stop creating software-driven obsolescence, Hussain said.
“It boils down to writing software so that we’re not forced to throw away things,” he said.
If a mobile phone, for example, is in working order but unable to run updated software the user needs, that doesn’t serve to reduce the digital carbon footprint, Hussain said.
6. Measure carbon intensity
In the business landscape, leaders treat growth in the arena of products and profits as gospel. That approach has had major environmental costs and is coming under scrutiny. But it highlights one issue around metrics: How do IT teams know if they’re doing better if the company is growing?
One of the challenges software developers face is that their industry is high growth, Hussain said. While they might be working to reduce their products’ emissions, their digital carbon footprint increases because more people are using what they produce. For example, a development team might have taken 15 different actions to reduce emissions, but the overall result is that their still responsible for a larger footprint because their organization’s sales continue to grow.
This is why it’s necessary to measure for intensity, Hussain said. The Green Software Foundation has created the Software Carbon Intensity (SCI) Specification that’s designed to score software systems on their carbon emissions, such as carbon-per-minute.
“With a software product, when [you’re measuring] carbon-per-minute, if you’re doing a good job it should be going down,” Hussain said.
7. Power the cloud with green energy
One important factor CIOs and their teams should consider more closely is the question of where they’re sourcing energy. Cloud software that draws its energy from windmills, solar farms or nuclear plants produces less carbon emissions than those that draw energy from coal or gas plants.
Cloud providers have different carbon footprints in different locations, Rogers said. IT leaders may want to revisit decisions about where best to house applications. Common wisdom has held that data centers should be close to users, but leaders may need to select locations farther away to address environmental sustainability issues.
“You could move [applications] to a lower carbon data center — which might reduce your cost as well — and it would reduce your carbon footprint,” Rogers said.
This may not be practical for organizations that need to deliver low latency services, such as gaming. But other applications, such as blogging sites, might not require the same rapid access.
“You can move the blog further away from users, and you can cut carbon and cut costs without really impacting your users’ experience,” Rogers said.
The challenge in this exercise is figuring out which data centers are more environmentally friendly, he said. “Finding information on an individual cloud provider’s data center sustainability [is difficult] and it’s hidden behind this wall of greenwashing to some degree.”
However, there are some cloud providers that aim for transparency, Rogers said. For example, Google Cloud publishes information on the carbon footprint of its data centers. IT leaders may also want to explore the Cloud Carbon Footprint project, an open source project designed to provide cloud carbon metrics for numerous cloud providers, for additional information, though Rogers specified he hasn’t verified its accuracy.
8. Understand renewable energy factors
Clean energy is better energy, but it’s not always so easy to access.
While it’s preferable for data centers to run on renewable energy, there’s only so much of it to go around, said Jason Jay, director of the Sustainability Initiative at MIT Sloan School of Management in Cambridge, Mass.
“Even if you’re incrementally causing [renewable energy infrastructure] to be built when it wouldn’t have otherwise been built,” Jay said, “that particular strip of offshore wind or that particular mass of solar is being taken to fuel your data centers and not the steel plant or the office building or the water treatment plant — or whatever other uses there are for electricity — who are now going to be on gray electricity.”
For organizations to become truly sustainable, they can’t rely only on buying their way out of the carbon emissions problem by purchasing renewable energy.
“Another challenge for CIOs is to learn how to become more sophisticated thinkers about the energy system,” he said.
This extends to public policy.
“To do some of these things right, ultimately public policy is going to be needed, and so the last piece of this challenge for CIOs is how to be politically engaged and how to help raise the tide for all boats, and not just be the one responsible leader,” Jay said.