INTRODUCING ECOMODE

REDUCING OUR DIGITAL CARBON FOOTPRINT

5 MINUTE READ. PUBLISHED 5 MAR 2023. UPDATED 19 JAN 2024.

CREATIVE COMMONS CC BY ELECTRO STRATEGY STUDIO. WRITTEN BY ADRIAN JARVIS. CONTRIBUTIONS FROM PRODUCT FOR NET ZERO.

TL;DR EcoMode is a concept for digital products & services that encourages users to choose app settings that reduce carbon emissions. It faces challenges to widespread implementation, especially convincing people to balance reduced functionality with increased sustainability.

The Challenge.

Achieving Net Zero in the Digital Age.

The fourth industrial revolution, fuelled by the rapid growth of digital technologies, has transformed every aspect of modern life. From healthcare to transportation, manufacturing to finance, technology has brought unparalleled efficiencies and conveniences.

However, every byte of data used by digital products and services has a carbon footprint. The accelerating carbon footprint of digital products poses a significant challenge to achieving Net Zero targets by 2030/50.

RATIONALE: Digital products and services are responsible for 4% of global greenhouse gas emissions - roughly the same amount as the aviation industry - and this carbon footprint is expected to grow at a rate of 10% per year. World Economic Forum Davos, 2022

And there are wider effects on global resources, created by our reliance on digital services, and their continued development:

EMERGING TECHNOLOGIES: Microsoft disclosed that its global water consumption spiked 34% from 2021 to 2022 (to nearly 1.7BN gallons), a sharp increase compared to previous years that outside researchers tie to its AI research. AP News, Sept 2023

FLOPS AND TOKENS: Artificial Intelligence is a known energy hog. Although precise energy usage varies between models, no one argues that the numbers are not sustainable. Information Week, Dec 2023

Reducing carbon emissions and other environmental impacts are becoming a priority; forcing us to rethink the way we build and use applications.

In 2020, Apple achieved carbon neutrality for its global corporate operations and announced Apple 2030: a bold strategy to be carbon neutral across its entire value chain by 2030. Apple, Sept 2023

However, under the GHG Protocol ‘Scope 3’ includes the impact of product use; emissions that do not directly come from manufacturing and servicing, that are upstream or downstream of the value chain.

IMPERATIVE: The environmental impact of using digital products is a critical, and often ignored, issue that can only be addressed though better software design and engineering.

DESIGNING LOW CARBON Software.

INTRODUCING ECOMODE.

One potential solution is the adoption of ‘EcoMode’; providing people with the ability to make more sustainable choices — without compromising on the core functionality of the apps and digital services they rely on.

SOLUTION: EcoMode is a user-initiated flag within a digital app or device that triggers low-carbon settings — for how services use data, media, and cloud services.

Incentivising developers to build digital products that prioritise sustainability — without restricting competitive feature design.

OUTCOME: EcoMode has potential to reduce the environmental impact of digital products and services. By enabling users to choose sustainable behaviours and guiding developers to design software to reduce carbon, EcoMode could contribute significantly to achieving global Net Zero targets.

Advancement on Low Power Mode.

EcoMode may be considered an advancement on Low Power Mode, an existing feature in both Android and Apple iOS. EcoMode would be a similar switch in the device settings, supported by developer guidelines for implementation.

COMPARISON: Low Power Mode primarily focuses on extending battery life by reducing power consumption. EcoMode goes a step further to actively minimise the device's carbon footprint by optimising wider runtime behaviour and use of cloud-based services.

By providing people with an option to reduce their carbon footprint, EcoMode is a simple and promising idea to promote carbon awareness and action — and a step in the right direction towards a net positive future for digital.

Barriers to Implementation.

We will need to overcome significant barriers to achieve the adoption necessary to reduce our digital carbon footprint, at scale.

  • INVESTMENT: changes to design requires purpose, time and budget.

  • MOTIVATION: how to encourage all stakeholders to adopt, use and support.

  • RESISTANCE: rejection of reduced functionality and perceived inconvenience.

  • IMPERATIVE: low priority, with a belief that other parties are responsible.

  • COMPATIBILITY: potential technical challenges to integrate into existing systems.

  • INCENTIVES: markets are driving Big Tech to focus on data-rich, AI-driven services.

Resources FOR Sustainable Digital Design.

The list of resources for sustainable design of digital products and services has been moved to its own page here (opens in a new tab).

APPENDIX.

Implementation OF ECOMODE.

Here are some potential ways of implementing EcoMode; to reduce carbon emissions while maintaining the functionality of digital products and services:

  • OS AND APP STORES: System-level switch that defaults to OFF, with user prompt and developer guidelines for implementation.

  • MOBILE: disabling push notifications, reducing the frequency of updates and limiting geolocation to reduce data usage.

  • SOCIAL MEDIA: manual refresh, data compression, and lower-resolution images and videos.

  • CONTENT STREAMING: adaptive resolution video and audio quality to reduce the data and energy.

  • E-COMMERCE: reduced-size image formats and disabling animations to optimise performance, and reduce carbon footprint.

  • PRODUCTIVITY AND CREATIVITY: compressing data, limiting cloud services and optimising use of artificial intelligence for assistive services.

  • CLOUD COMPUTING: hosts run on renewables, more effective resource allocation and virtual machines.

POTENTIAL FOR MAJOR PLATFORMS.

All the major OS and platforms could implement changes to the runtime of their products, based on how a user sets EcoMode on their device.

Shown here are some examples of routines that could have an effect on reducing the carbon footprint of key digital service (work-in-progress, suggestions welcome).

OPERATING SYSTEMS

  • GOOGLE ANDROID: Battery Saver, Adaptive Battery, Adaptive Brightness, Data Saver, Restrict Background Processes

  • APPLE iOS: Low Power Mode, Reduced Motion, Reduced Transparency, Optimized Battery Charging, Background App Refresh

SOCIAL MEDIA PLATFORMS

  • FACEBOOK / INSTAGRAM: Video/Audio Compression, Dark Mode, Data Saver

  • LINKEDIN: Optimized Images, Reduce HTTP Requests, Lazy Loading

  • PINTEREST: Low-Power Mode, Reduce Motion, Optimized Images

  • SNAPCHAT: Video/Audio Compression, Dark Mode, Data Saver

  • TIKTOK: Optimize Video/Audio Compression, Reduced Frame Rate, Video/Audio Quality Settings

  • TWITCH: Video/Audio Compression, Data Saver

  • YOUTUBE: Video/Audio Compression, Reduced Video Resolution

COMMUNICATIONS PLATFORMS

  • WHATSAPP: Limit Data Usage, Reduce Video/Audio Resolution, Media Download Settings

  • ZOOM: Optimize for Low Bandwidth, Disable HD Video, Reduce Frame Rate, Audio Only

CONTENT PLATFORMS

  • APPLE MUSIC: Audio compression, Limit Cloud-based Algorithms and Processes

  • NETFLIX: Video/Audio Compression

  • SPOTIFY: Audio Compression

OTHER DIGITAL SERVICES

  • ADOBE CC: Power Saving Mode, Dark Interface, Optimized Rendering, Energy-Efficient Coding

  • AMAZON WEB SERVICES: Amazon EC2 Auto Scaling, Compute Optimizer, Resource Groups

  • GOOGLE CLOUD: Carbon-Aware Computing, Resource Optimization

  • MICROSOFT AZURE: Sustainability Calculator, Power Calculator, Azure Advisor, Virtual Machines

  • OPEN AI: Optimize AI Models for Efficiency, User Initiated Prompts

EXAMPLE Variables FOR Reducing Carbon.

Shown here are potential ways to adjust global and local variables to optimise the efficiency of software systems — to reduce the impact of software and websites.

  • Image, Video and Audio Compression: Set values for lower fidelity media eg. video at 240p vs 4K

  • Manual Refresh and Sync: State to turn ON/OFF auto-refresh eg. Default OFF. Pull to refresh

  • Data Compression / Limitations: Automatically compress data for transfer eg. LZx libraries

  • Low-Power and Sleep-Modes: State to turn off background processes / sleep eg. Default ON. Settings override.

  • Dark Mode / Brightness: UI in Dark Mode; low brightness palette choice eg. Default ON. Settings override.

  • Location-Based Services: Restrict LBS to only those essential for operations initiated by the user, such as mapping eg. Default CRITICAL ONLY

  • Push Notifications: Limit use of unnecessary notifications, especially from remote servers eg. Default CRITICAL ONLY

  • Local Storage (vs Cloud Storage): Limit unnecessary cloud storage and retrieval eg. Default LOCAL. Backup DAILY

  • On-Device Processing (vs. Cloud) and Client-Side Rendering (vs. Server-Side): Selection of most energy efficient processing.  This may not be locally.  Consider data transfer. eg. Default LOCAL / EDGE / CLIENT

  • Cached Content / Edge and Long-Term: Maximize caching to limit new data requests eg. Flexible cache with user set limit

  • Content Delivery Networks (CDNs): Optimize for local servers and cached content eg. Status code 200

  • Limited and NoSQL Database Queries: NoSQL databases are more efficient due to how they deal with unstructured data eg. Optimized queries, caching and load balancing

  • Vector Graphics (vs. Raster): Graphics formats that use less data. They also speed up apps eg. SVG over PNG or TIF

  • Progressive Content Loading: Load content necessary for page interaction eg. server-side e.g. Flask or Django

  • Compute for AI Models and Algorithms: Limit the load and quality of AI queries eg. Limit tokens for generative AI

  • Settings for Encryption, Hashing, Sorting and Search Algorithms: Improve efficiency and resource optimization, while balancing data size and security eg. Libraries with flexible and more efficient algorithms

  • Protocols for Network and Hypertext (HTTP/2 and 3): Protocols that reduce the number of connections and calls made by applications eg. Multiplexing, Header Compression, Server Push

  • Use of Virtualization, Containerisation and Microservices: Potentially reduce the energy consumption and carbon footprint associated with running data centres and deploying applications eg. Optimize resource usage, simplify maintenance, and cloud services that run on renewables