8 Sustainability

Sustainability in software means looking beyond readability, simplicity, and failure handling to consider the resources software consumes: CPU, memory, storage, network traffic, electricity, and hardware. The chapter frames sustainability through environmental, social, and economic pillars, with a focus on the environmental impact of digital technology. Software contributes to emissions through the electricity used by devices, networks, and data centers, but also through the hidden lifecycle cost of hardware manufacturing, transport, and disposal. Because much of a device’s carbon footprint exists before it is ever used, extending hardware lifespan and avoiding software-driven obsolescence are essential sustainability goals.

The chapter presents sustainable coding as a practical extension of software craftsmanship. Clean code concerns maintainability, performance concerns speed and responsiveness, and sustainable coding concerns impact: using the smallest reasonable amount of infrastructure and energy while still meeting user needs. The main levers are reducing hardware demand and lowering electricity consumption, especially when power comes from carbon-intensive sources. To help developers act on these levers, the chapter introduces green design patterns: frugality, lean packaging, lean storage, lean communication, efficient execution, memory efficiency, and carbon awareness. Together, these patterns encourage developers to build only valuable features, remove unnecessary dependencies and dead code, store and transmit less data, optimize computation, manage memory carefully, and shift heavy work to cleaner times, regions, or modes of operation.

The chapter also explains that measuring the exact carbon footprint of software is difficult, because emissions depend on uncertain factors such as electricity sources, deployment regions, user devices, and embodied hardware carbon. Still, estimation methods such as Software Carbon Intensity can provide useful approximations once software is running. Developers can also use frontend analysis tools, static analysis, benchmarks, profilers, memory inspection tools, and carefully applied AI-assisted reviews to detect wasteful patterns earlier in the development process. The central message is that sustainable coding is not about optimizing everything blindly; it is about making reasonable, evidence-based choices that reduce waste, lower costs, preserve maintainability, and build software that respects both users and the environment.

How sustainable coding fits with clean code and performance.

Using the two levers of sustainability: electricity and hardware usage.

Green design patterns

• Sustainability is based on three pillars: environmental, social, and economic. These aim to balance long-term economic profit with reduced ecological impact and improved well-being for people.
• As of 2025, information and communication technology (ICT) accounts for about 4% of global greenhouse gas emissions. This impact is caused by both the electricity consumed by digital systems and the carbon-intensive manufacturing of their hardware.
• A carbon footprint measures the total greenhouse gases produced by an activity, expressed in CO₂ equivalent. This converts different gases such as carbon dioxide, methane, and nitrous oxide into a single comparable unit.
• The carbon footprint of software from electricity usage depends on both the amount of electricity consumed and how carbon-intensive the electricity is in the region where it runs.
• The embodied carbon footprint is the total emissions generated during the manufacturing, transport, and end-of-life processing of a device.
• Technological obsolescence occurs when functioning hardware becomes unusable, either when software support ends or when new software requirements exceed device capabilities.
• Efficiency is the ability to achieve acceptable performance while using as few hardware resources as possible.
• Green design patterns are practical strategies that reduce the environmental impact of software by acting on two levers, lowering electricity consumption and limiting resource usage. There are seven patterns:
• The frugality pattern keeps systems simple and resource-efficient by building only the features and technical elements that deliver real value to clients and users.
• The lean packaging pattern removes unused dependencies and dead code to reduce package size and transfer cost.
• The lean storage pattern limits stored data and removes accumulated operational waste to reduce storage.
• The lean communication pattern lowers network traffic by reducing call frequency and shrinking payloads.
• The efficient execution pattern reduces CPU work through better algorithms and workload smoothing.
• The memory efficiency pattern minimizes memory usage by choosing lightweight structures and controlling cache growth.
• The carbon-aware pattern reduces emissions by scheduling work during cleaner energy periods, shifting computation to lower-carbon regions, and scaling resource usage down when electricity is more carbon-intensive.
• Tools can only provide an estimate, and the most reliable estimates appear once the software is deployed, whether through cloud provider dashboards or installed monitoring software.
• Front-end carbon footprint can be analyzed with webpage analysis tools that measure page weight and network requests, as well as static analysis tools that flag wasteful coding patterns.
• Back-end impact can be analyzed with static code analysis, memory management tools, and performance profilers.

FAQ