Insights
Calculating the carbon emissions of our cloud computing
Like many businesses, we’re becoming increasingly conscious of the impact we're having on the environment. And while our primary focus is always going to be on safeguarding children against online crimes, we also want to lessen our impact on the environment as much as possible.
We’re early on in this process, but we’ve recently taken an essential first step: to measure the carbon emissions of our cloud computing, and to create a plan for reducing them. Here’s how we’re doing this with the support of Tailpipe – whose platform can calculate the emissions of cloud computing – and GreenOps consulting partner, DigitalPebble.
Understanding the carbon cost of our cloud computing
We rely on cloud computing heavily here at CameraForensics. It’s what allows us to efficiently process images and data on a huge scale – so that we can develop novel online imaging techniques and digital investigative solutions – without us needing to invest in our own computing infrastructure.
Using Tailpipe, we’ve started to measure how much carbon we generate through these activities. Tailpipe looks at the billing and usage information from our cloud services, primarily hosted with AWS, and estimates the energy used by each, taking into account the type of hardware used. Then, it combines this with data from the location in which those services are running, considering the efficiency of its data centres and the grid mix of the region it’s based in.
Once we got the data from Tailpipe, DigitalPebble provided a useful analysis by breaking down the emissions per service and operations. We were able to calculate our:
- Operational emissions – the carbon emissions generated from our day-to-day cloud computing activity.
- Embodied emissions – the emissions generated by the entire lifecycle of our cloud computing infrastructure, from manufacturing to shipping and to final disposal.
We learned that each month, we generate the equivalent of 1.3 tonnes of CO2 because of our use of cloud computing. To put this into perspective, this is roughly the same as:
- Driving a petrol car 4756 miles
- Driving from the top to the bottom of South America
- 1.4 people taking a return flight from London to New York.
How we plan to reduce our associated emissions
After analysing the data, DigitalPebble recommended three ways that we could start to reduce our cloud computing emissions.
1. Move our cloud operations to less carbon-intensive regions
DigitalPebble advised us that the biggest step we can take to reduce our cloud carbon emissions is to move our cloud operations to a country with a less carbon-intensive grid mix. To put it simply, a country that relies more on renewable energy than coal, oil or gas.
Currently, most of our cloud operations are hosted on servers in the US which has a carbon intensity of 375g CO2eq per kWh on average. This is due to an energy mix consisting mainly of gas and coal. By migrating them to somewhere like the Sweden AWS region, which has an average carbon intensity of about 20g CO2eq, we could reduce our associated carbon footprint to one-twentieth of its current size!
This isn’t going to be an overnight process for us, and there are other factors which will determine where certain infrastructure and data need to be located. However, we have now committed to increasing our use of servers in low-carbon regions, such as Sweden, particularly when deploying the infrastructure needed for our R&D work – which is often carbon-intensive.
2. Use more energy-efficient servers
Using more energy-efficient instance types is another way to reduce how much carbon our cloud computing generates. By using the most appropriate instance types for our activities, we can make sure we’re not using more CPUs or GPUs than needed.
Fortunately, we already have good measures in place to assess utilisation – AWS provides some basic but effective tools to help here – and we also make very good use of cost allocation tags to help us understand what parts of our business are most heavily using cloud resources.
3. Understand our wider environmental impact
DigitalPebble is also helping us to understand how our cloud computing emissions compare to the rest of our emissions. For instance, those we generate by flying internationally.
As we begin our carbon reduction journey, focusing on our cloud computing is the right place for us to start. We have the data and recommendations ready to put into action, and doing so won’t compromise on key aspects of our work – like attending conferences with international law enforcement agencies, for example.
Nevertheless, with this broader understanding of our environmental impact, we’re in a good position to reduce our carbon emissions further in the future.
We’re grateful to Tailpipe and DigitalPebble for supporting us on this journey so far. If you rely on cloud computing like we do, we truly recommend working with them to measure your associated carbon emissions, and get tailored recommendations for reducing them.
