Value judgments and sustainable computing
I often see people making value judgements when discussing sustainable computing. This puts a higher value on one type of computer use versus another. The implication is that we should do less of some things because their value cannot justify the environmental impact.
These judgments are commonly expressed in terms of waste: duplicate photos are a waste, social media is a waste, “hoarding” data is a waste. The analogy is between bits and atoms where the atoms (in the form of physical objects) take up space that could be used for something else (supposedly better), or are discarded and cause disposal or recycling problems.
This analogy fails because it misrepresents the fundamental innovation of digital services: zero marginal cost and abundance. Bits and atoms are not the same. At the scale of the internet, storing additional data does not deprive someone else of their ability to store data. Streaming video at 4k does not prevent someone else from also streaming video at 4k. This is completely different from the physical world where there are real limits and space is truly “used up”.
Another analogy is with harmful or addictive substances, like alcohol. This is the idea that we’re “drunk” on computing resources like cloud storage or data, and we must adopt the idea of “digital sobriety”. One supposed solution is an “internet speed limit” to stop wasteful behavior. This is often used in the context of data center or network energy, particularly when data centers are competing for physical resources like water.
These arbitrary rules are applied by people who feel like they have the right to judge the behavior of others. How the rules should be define are ambigious, or are a rehashed argument from authority. They distract from the real challenges of building sustainable infrastructure by focusing on inconsequential actions like deleting unwanted photos or reducing video streaming quality settings. These are meaningless interventions that confer a false sense of “making a difference” that could be better directed elsewhere. Or worse, exacerbate the problem.
Effort is better directed at altering systems so that everyone can benefit from improvements at scale. This means working towards greater deployment of clean energy systems that power our IT. It means increasing transparency to help analyze the energy consumption of cloud computing, so we can understand the energy (and carbon, water, materials..) footprint of services built on top.
The goal of sustainable computing is to be able to use more and more IT services whilst reducing the environmental impact. This means providing data, tools, dashboards, and best practices for those developing and operating the systems we use every day. It’s already happening in networking with data transfer volumes vastly increasing at the same time as energy consumption decreases.
Puritan value judgments don’t help and any analysis based on them should be rejected.