...

Cloud Sustainability

Dimitur Stefanov
June 5, 2024
Read: 5 min

This article on cloud sustainability is a part of a series on carbon analytics published on the Infinite Lambda Blog.

Appreciating the cloud is becoming more and more like appreciating electricity. Don’t get me wrong — the cloud still has a lot of catching up to do before it can be considered equally essential. However, the more you think about it, the more you realise that it is at every corner and behind every screen.

For ordinary people, this includes services like video and music streaming, online banking and cloud storage. Companies, on their part, increasingly rely on cloud providers to store and make sense of their data, host applications and keep their customer base secure, among other things.

Needless to say, these services do not come out of thin air. Commercial and private cloud data centres, filled with stacks of hardware and miles of cables, are dotted around the world. Each centre is supported by staff and systems, ensuring continuous power supply, appropriate temperature and humidity, technical maintenance and security. The more critical a data centre is, the more sophisticated these support systems are.

Cloud data centres’ carbon footprint

The neat image I described above has a darker side, being a significant contributor to greenhouse gas (GHG) emissions. Data centres account for 1% of the global energy consumption; in the US alone, their electricity demand constitutes 2% of national energy consumption.

According to a study by Charlotte Freitag and colleagues, information communication technology produces between 1.8% and 2.8% of global GHG emissions and data centres account for between 18% and 41% of those.

It is tempting to brush off data centres’ contribution to GHG emissions as something laughably obvious: they run on electricity, power generation produces emissions, nothing new here. However, the reality is multifaceted.

Data centres’ emissions

Data centres’ emissions are often attributed to:

  • Hardware electricity consumption: just like your laptop or phone, servers use energy, even when on stand-by. As demand for servers’ resources grows, so does the electricity demand.
  • Operational electricity consumption: these are also somewhat obvious. Think of all the systems in place that ensure an appropriate temperature, humidity, lighting and security that need to be powered somehow.
  • Carbon intensity: a unit of electricity will have a different ecological impact depending on the mix of fuels used for its generation. Data centres with identical energy needs may have significantly different carbon footprints depending on location, energy provider, fuel availability and even time of day.

Data centres also contribute to GHG emissions indirectly, something that can be easily overlooked. Examples include:

  • Hardware production: the various components were once raw materials in the ground. Extraction, manufacturing and recycling have significant energy demands, meaning that a piece of hardware has impacted our climate long before it gets used. This is known as embodied emissions.
  • Data centre logistics: delivering necessary equipment will naturally increase indirect emissions, the most obvious factors being the method of transportation and fuel used. Warehousing new equipment and disposing of waste contribute to the carbon footprint in their own right as well.
  • Staff operations: data centres may require a whole array of engineering, security, cleaning, catering and managerial staff to run properly. All of these people will impact the centre’s carbon footprint by virtue of their commute, accommodation and the goods and services they need to do their jobs.

Cloud sustainability: data centres' energy consumption

Beyond data centres: assessing cloud sustainability

We know that:

  1. We increasingly rely on data centres and the cloud because they are cool;
  2. Using these makes our planet less cool for a variety of reasons.

Now comes the part where we list your favourite cloud-based activities and the environmental bill they incur.

Disclaimer: Get ready to feel guilty for sending unnecessary queries to your database while an episode of Friends is playing in the background.

Ready?

Unfortunately, it is not that simple. Cloud sustainability and the impact of your usage depends on numerous factors. In fact, the variables are so many that providing a single emissions factor for a unit of some vaguely defined activity like streaming or asking ChatGPT a question is inaccurate at best and counter-productive at worst.

The issue with reliable estimates

Let’s consider video streaming. You have likely come across headlines warning you that an hour of Netflix is equivalent to a short drive with your car. While nobody is denying that streaming has a carbon footprint, these headlines are misleading because “actual [energy] consumption depends on the device, network, connection, and resolution”, as explained here.

On device level alone, we have incredible room for variability. A phone and a flat-screen TV will differ significantly in embodied emissions, energy demand and the carbon intensity of the electricity used to power them.

If we cannot provide a reliable emissions estimate for one hour of streaming, what about companies’ cloud usage? The head-spinning galaxy of providers, services, configurations and usage patterns leaves us with no other option than doing these estimations on a case-by-case basis.

Outpacing cloud sustainability efforts

For argument’s sake, let’s assume we can come up with a reliable emissions factor for each cloud-based activity. So what? The world around us is changing so fast that within a few years these factors will likely be obsolete.

On the supplier side, data centres and devices are becoming ever more efficient, manufacturing takes less energy and raw materials than before, companies are becoming more sustainability conscious. Even at the cutting edge of AI, researchers are looking for model training methods that require less resources without sacrificing performance.

The consumer side is no different. More and more ordinary people have regular access to streaming, social media, AI and other cloud-based services than ever before. Data-driven companies are increasingly relying on cloud providers and are now rushing to implement AI solutions to their business problems. This ongoing increase in demand can easily outpace advances in sustainability and efficiency, raising questions about providers’ and governments’ response.

Cloud sustainability and carbon analytics

Where does that leave carbon analytics?

Although we only had a high-level glimpse at the complexity of assessing cloud sustainability and environmental impact, you get the idea — it depends on a multitude of variables. However, the presence of these should not discourage us. Think about what world you would rather live in. Is it one where the carbon emissions of your data pipeline or Spotify subscription are more or less fixed, meaning that the only way to reduce your footprint is to limit your usage? Or is it a world where these activities come with knobs and switches you can use to optimise and ultimately reduce your emissions while meeting your needs?

When considering cloud sustainability, estimating the carbon footprint of your company’s usage, or any other activity for that matter, should not be framed as a project that can be ticked off and forgotten once some arbitrary acceptance criteria are met.

Instead, carbon analytics should be thought of as an ongoing process akin to a company’s financial or HR processes. One where you lay the foundations and iteratively revisit and build on top of them to keep up with the dynamic nature of your business and the world around it.

Such a process should not exist in a vacuum but is instead based on coordination among key company departments. And most importantly, the people engaging in carbon analytics should embrace opportunities to learn from others and share their experience with the rest of the world.

If you are interested in what implementing carbon analytics could look like in your organisation, reach out. The Infinite Lambda experts are ready to talk to you about your specific goals and how you can address them by leveraging the analytic capabilities that the modern data stack has unlocked.

More on the topic

Everything we know, we are happy to share. Head to the blog to see how we leverage the tech.

Tag-Based Masking in Snowflake
Tag-Based Masking in Snowflake: Practical Guide with Scalable Implementation
As data continues to be a critical asset for organisations across industries, safeguarding sensitive information while enabling data access for authorised users is a constant...
June 11, 2024
How to measure happiness and safety in tech teams
How to Measure Happiness and Safety in Tech Teams
Software product development initiatives can go awry for a whole range of reasons. However, the main ones tend not to be technical at all. Rather,...
May 30, 2024
why sustainability analytics
Why Sustainability Analytics
We all like a sunny day. Kicking back in the garden with the shades on, cool drink in hand and hopefully a liberal amount of...
May 8, 2024
Data diff validation in a blue green deployment: how to guide
Data Diff Validation in Blue-Green Deployments
During a blue-green deployment, there are discrepancies between environments that we need to address to ensure data integrity. This calls for an effective data diff...
January 31, 2024
GDPR & Data Governance in Tech
GDPR & Data Governance in Tech
The increasing focus on data protection and privacy in the digital age is a response to the rapid advancements in technology and the widespread collection,...
January 18, 2024
Data masking on Snowflake using data contracts
Automated Data Masking on Snowflake Using Data Contracts
As digital data is growing exponentially, safeguarding sensitive information is more important than ever. Compliance with strict regulatory frameworks, such as the European Union’s General...
January 17, 2024

Everything we know, we are happy to share. Head to the blog to see how we leverage the tech.