Carbon-Efficient Software Design and Development: A Systematic Literature Review

TL;DR

This paper tackles the problem of carbon Efficiency in software by conducting a systematic literature review of state-of-the-art proposals for carbon-efficient software design and development. It identifies 65 primary studies and classifies them using a 5W1H taxonomy (Who, What, Why, Where, When, How) to reveal patterns across sustainability guidelines, reference models, measurement solutions, and software-improvement techniques. The study finds a predominance of generic, low-TRL reference models and guidelines, with far fewer efforts directly minimizing carbon and a heavy focus on software developers as stakeholders; energy consumption is often used as a proxy for carbon, highlighting the need to incorporate energy mixes and direct carbon accounting. The authors propose seven research directions to standardize definitions, broaden stakeholder involvement (including architects and sustainability engineers), and increase TRL and domain-specific guidance, aiming to bridge the gap between research and industry practice and accelerate real-world adoption of carbon-efficient software design. The work provides a foundational map for researchers and practitioners to select applicable guidelines and tools, and to pursue domain-aware, lifecycle-spanning approaches that better align software development with environmental goals.

Abstract

The ICT sector, responsible for 2% of global carbon emissions, is under scrutiny calling for methodologies and tools to design and develop software in an environmentally sustainable-by-design manner. However, the software engineering solutions for designing and developing carbon-efficient software are currently scattered over multiple different pieces of literature, which makes it difficult to consult the body of knowledge on the topic. In this article, we precisely conduct a systematic literature review on state-of-the-art proposals for designing and developing carbon-efficient software. We identify and analyse 65 primary studies by classifying them through a taxonomy aimed at answering the 5W1H questions of carbon-efficient software design and development. We first provide a reasoned overview and discussion of the existing guidelines, reference models, measurement solutions and techniques for measuring, reducing, or minimising the carbon footprint of software. Ultimately, we identify open challenges and research gaps, offering insights for future work in this field.

Figures (11)

• Figure 1: SLR process for identifying and classifying existing literature of interest.
• Figure 2: Frequency of stakeholders over the selected studies. Since some studies target multiple stakeholders, the sum of frequencies is higher than the number of selected studies.
• Figure 3: Frequency of proposals in the selected studies (with MS=measurement solutions, RM=reference models, SG=sustainability guidelines, and SIT=software improvement techniques).
• Figure 4: Coverage of (a) source code availability and (b) TRL in the selected studies.
• Figure 5: Frequency of energy- and carbon-related goals in the selected studies, distinguished by proposal. Since some studies pursue multiple goals, the sum of frequencies is higher than the number of selected studies.