Ok, well not totally meaningless. But it certainly isn’t always the untouchable source of truth that is often perceived to be. My name is Otis Laundon, and throughout my Green Web Platforms internship I have focused so much on trying to find data. Data on energy usage, energy grid carbon intensity, and equipment manufacturing, to be ultimately used as the ingredients to cook up an estimate of carbon emissions. But the more data I find, the less answers it seems to provide. 

In an area such as digital sustainability where much of the data is uncertain, many subjective decisions have to be made as to which figures to use.  

Example – Hard Drives 

If we wanted to figure out what the environmental impact of data storage is, part of that would come from the lifecycle of the digital storage hardware, i.e. the hard drives. Since this is happening in the cloud on computers that we will never touch, see, or even know exactly what they are, many aspects have to be estimated or assumed.  

They could be 1 TB or 10TB drives, they might consume anything from 4 to 10 watts of energy, they could be replaced after anywhere between 2 and 6 years, etc. Let’s use these numbers in an example calculation for the impact of storing 1 GB for 1 year in the cloud. 

We will assume embodied emissions of 20kg co2e per TB hard drive capacity [1], and use the formula 

Emissions per GB year = (embodied_emissions_per_harddrive / lifetime + operational_emissions_per_harddrive_year * grid_intensity) / capacity 

Which using the range of numbers above could be as low as 

(20 / 6 + 0.004 * 365 * 24 * 0.125) / 10000 = 0.00077 kg CO2e / GB year, 

or as high as 

(20 / 2 + 0.01 * 365 * 24 * 0. 125) / 1000 = 0.021 kg CO2e / GB year. 

That’s a difference of over 27 times! And highlights how large the range of plausible values are for just a small part of the overall environmental impact of a web service, its storage. 

The final estimate of “emissions from storing a GB for a year” is more a product of my assumptions than any concrete fact. 

So what’s the point? 

My point is that an entirely data-driven approach to reducing the impact of digital is doomed to fail. We just don’t have the data available in most cases. That being said, data can tell a story and can definitely play a role in bringing about change. If we are consistent about how we do the calculations using the data, we can see when we make a negative or hopefully positive change. It also makes communication of the impact of services on the environment easier. 

I believe that the role of sustainable web design best-practices such as the W3C’s Web Sustainability Guidelines is as important as quantifying the impact itself. Design best-practices provide a direction for action, just as data promises to, but when our data is so uncertain it would be irresponsible to follow it blindly, chasing lower estimates for emissions rather than real change. 

For most things that websites aim to achieve, we have more than enough computing power. We can already transfer information across the world in real time. That is a phenomenal achievement. I want to see a shift to refining and stripping back software rather than constantly cramming more features into it, favoring simpler ways to meet our needs, and not trying to grow the beast that is the internet just for growth’s sake. 

Sources 

[1] : The Dirty Secret of SSDs: Embodied Carbon 

Data is Meaningless / Digital Learning Applications and Media by blogadmin is licensed under a Creative Commons Attribution CC BY 3.0

Share