Everything that we create and share on the internet is stored somewhere. However, most of us aren’t thinking about how and where this data is stored. As developers, we may think that our data is ephemeral, but all of it lives somewhere. When we work with code all day, it can be easy to forget about all of the physical things that are powering our world.
The majority of all data in computing is stored in a data center. A data center is one or many buildings that store servers and all of the hardware associated with them. Before data centers were ubiquitous, we had to store our servers on our own, sometimes in a room. But people found that data centers were easier to use than maintaining their own server rooms, and they became much more common. And if you use a cloud provider, then you are using the data centers that they provide.
Cloud providers are companies that rent out servers in their data centers. Storing content in a cloud provider’s data center is more efficient than running a data center of your own. You don’t have to maintain your own machines, or your own network. You have access to better hardware and improved security infrastructure, like backups and distributed databases. When you sign up to use a cloud provider, they will promise you little to no downtime. In other words, they say that they will keep things running.
But what do providers have to do in order to ensure operability, and no downtime? In order to make sure that you can always access your data and in order to keep their systems reliable, they often have more than one instance of a server in different locations. This is called redundancy. But the more servers you have, the more servers you need to power and maintain. And servers are demanding! A server requires energy to be powered, which comes in the form of electricity. But servers also use a lot of energy and emit it in the form of heat. If a server emits too much heat, it will overheat and fail. So, the server also has to be cooled! Because of these hardware constraints, every data center has some mechanism to remove hot air and supply cool air. So, not only do we need energy to power the servers, we need energy to cool them down, too.
But just how much energy are we talking about? By conservative estimates, data centers worldwide use approximately 200 terawatt hours (TWh) a year. One terawatt-hour is equivalent to a trillion watts of energy consumed in one hour. For some perspective, Scotland has a population of over 5 million and requires only 25 TWh of electrical energy each year. Overall, data centers currently demand somewhere between 1-3% of the world’s global electricity.
If data centers are using so much of the global electricity supply, it’s worth asking: where does the energy from? The electricity to power data centers comes from the local power grid. Unfortunately, many data centers are relying on power grids that use fossil fuels as their main power source. Fossil fuels emit greenhouse gasses into the atmosphere. Data suggests that the IT and communication technology industry accounts for approximately 2% of greenhouse gas emissions. Out of this 2% of emissions, conservative estimates suggest that data centers are responsible for at least 0.3% of that. Another way of thinking about this is that the tech sector has a carbon footprint equivalent to the airline industry.
So what are the cloud providers are doing about this? Well, it depends a lot on the provider and their values. There’s a great white paper that was published last year called "The State of Data Center Energy Use in 2018". It was written by two folks in the industry, Paul Johnston and Anne Currie. This paper goes into a lot of detail about the environmental impacts of six of the biggest cloud providers, and it actually gives them a grade in terms of the sustainability of their servers! I’m only going to talk about two of these today, but I encourage you to check out their entire white paper, it has some wonderful resources.
We’ll start with Amazon's cloud provider service, AWS, which is the largest cloud provider. They have a commitment to be 100% renewable by 2030. Amazon allows you to store your server instances in different zones across the globe. They have 25 zones, but only 5 of them are carbon neutral. If you store your data in a carbon neutral zone, AWS will purchase something called a "carbon offset". When AWS buys an offset, they are purchasing renewable energy and putting it back into the electrical grid. These offsets are what makes these zones "carbon neutral". It’s important to note that if you store your data in one of these zones, you’re still emitting greenhouse gasses. You’re just slowing down your emissions by buying more renewable power generation from elsewhere. Offsets are not a long-term solution because they don’t remove carbon emissions from the atmosphere, they just slow it down.
Even though AWS offers 5 carbon neutral regions, it also has regions like US-EAST-1, in Northern Virginia, in the United States. A lot of companies host their data in this region. Unfortunately, Dominion Energy, the power supplier for this region, has doubled down on fossil fuels. As a result, this means that all of the data centers in this state are contributing to carbon emissions, because the power supplier is the one deciding what kind of energy source to use to power the electrical grid. And remember that goal that AWS had of being "100% renewable"? Well, despite this goal, they are continuing to open new data centers in this region. Since 2017, they’ve actually increased their operations in US-East-1 by 59%. On top of that, AWS is also not transparent about publicly reporting data on their current energy use, or how quickly their energy use is growing. This makes it hard to know if the renewable energy that they are buying through carbon offsets is anywhere close to offsetting how much energy they are actually using.
Now, if we take a look at Google, things get a bit brighter. The Google Cloud Platform buys carbon offsets for all of its servers. Their parent company, Alphabet, is the largest corporate buyer of renewable energy. Compared to AWS, Google has been pretty transparent about the fact that they actually can’t power 100% of their servers through entirely renewable energy! So, instead, they have adopted a different strategy. For each kilowatt-hour of energy that their servers consume, Google purchases a matching kilowatt-hour of clean, renewable energy and adds it back into the power grid. So, even though hosting on Google cloud causes new carbon emissions, 100% of those emissions is offset. Google is actually the leader in this sector, and is doing better than most other cloud providers. Microsoft is the only other provider that has actually met its 100% sustainability goal, which you can read about on the whitepaper.
Now that we know the reality of how these cloud providers are actually powering the data centers that house our content, the question is: how will the energy usage of data centers scale into the future? Multiple research papers have found that global data traffic is actually growing quite quickly. This means that the usage of data centers is going to grow, too. Over time, this will demand more energy. Here is the "expected case" projection for how the IT industry will grow over time. The estimate is that our industry will use somewhere between 8-21% of the total global electricity demand by the end of this decade. Given that we’re an industry that cares so much about efficiency and scalability, we need to start seriously thinking about this problem. I have some good news though: some of us are!
People have started solving this problem. Some companies are building their new data centers in colder climates, where it doesn’t take as much energy to cool servers. Others are recovering and recycling wasted heat from servers and re-using it in sustainable ways. Companies like Stripe and Basecamp, which don’t even have data centers of their own, have dedicated themselves to becoming fully carbon-neutral or carbon negative.
Each of us has agency in solving this problem, too. We can find out where our data lives, and whether it is stored in a region that is green and carbon neutral and powered by renewable energy, or whether is it contributing to the planet’s greenhouse gas emissions? There’s a great website called the Green Web Foundation that can help you answer this question when it comes to your own projects, as well as the apps that you might use every day. If your data doesn’t live in a green region, figure out what it would take to migrate your data to a different location or provider that is carbon neutral.
If you’re in the lucky position to start fresh and are provisioning new servers or databases, you can choose a clean cloud provider from the get go. Another great step you can take is to draw attention to this issue and talk about it with your team. If you’re at a small company, this might mean just talking about your cloud provider choices internally. But if you work at a big company, especially one that has a large enterprise account, you can pressure your cloud provider to release data and be transparent about where their energy comes from for powering their data centers.
And if you work at a company that provides cloud services, and if you have the power and the privilege to do so, you can do even more. There are some very active employees at Amazon that are doing exactly this. You can even build something to make it easy for people to find information about this topic.
It’s very hard to find data around energy consumption and these cloud providers, so it’s important to make this data accessible to everyone in the industry. A great example of this is the Cloud Sustainability Console chrome extension. It’s a simple but useful browser extension that highlights which regions in AWS are carbon neutral. But, if you take nothing else away from this, I hope the one thing that you will all do is be aware. Be aware of the physical impacts of what we do every day (even though we can’t always see them). Be aware of the impact that our industry is making on all of us, and on our planet. Our industry relies upon a finite resource, even if we can’t see it or even if we don’t think about it all the time. It’s our responsibility to take the first step and learn about this stuff. As I’ve learned more about how we store data, I’ve also realized that most of us don’t even know what this information will cost us down the road. We don’t yet know what the cost of data will be to us, or our planet.
But despite this, I don’t feel entirely disheartened. It might seem like all we’re doing as an industry is contributing to the problem, but there’s another way to look at it, too. We could be the ones who start to change it. Historically speaking, the field of technology has pushed the needle forward. I’m optimistic that we can do it again here, too, and set an example for other industries. We are a unique group because we aren’t just consumers of technology: we create it. Which means that we each have the power to influence and guide our industry, too. I hope to see all of us guide it in the right direction.