Sustainable Software Engineering

• Sustainable Software Engineering is a discipline that integrates environmental, economic, social, technical, and human dimensions into the full software lifecycle.
• It emphasizes measurable benchmarks like energy consumption, carbon footprint, and perdurability to balance performance with long-term sustainability.
• It incorporates agile, DevOps practices, and organizational strategies to embed sustainability standards and interdisciplinary frameworks in every development phase.

Sustainable Software Engineering (SSE) encompasses the principles, practices, and organizational structures required to design, develop, and maintain software that meets both present and future needs without compromising environmental, social, economic, technical, or human well-being. SSE incorporates system-wide and longitudinal concerns, situating sustainability as both a non-functional requirement and an operational imperative across the software lifecycle. The discipline is increasingly informed by standardized metrics, interdisciplinary frameworks, and empirical research highlighting the necessity of integrating sustainability into software engineering processes, education, and management structures.

1. Concepts and Dimensions of Sustainable Software Engineering

SSE is characterized by the integration of multiple sustainability dimensions into both product and process. Classical "pillar" models enumerate environmental, economic, social, technical, and individual dimensions, but recent theory argues that sustainability is both stratified (manifesting uniquely at individual, organizational, industry, and societal levels) and multisystemic (emerging from interactions among sociotechnical, economic, and natural systems) (McGuire et al., 2023).

Stratified and Multisystemic Perspectives

• Stratification: Each pillar's meaning varies by abstraction layer—for example, ecological sustainability ranges from component-level energy efficiency to ecosystem-wide climate impacts; economic sustainability may mean profitability for organizations and affordability for users.
• Multisystemic Nature: Sustainability outcomes are determined by both direct software attributes and the combined effects of legal, infrastructural, organizational, and psychological systems.

A conceptual model, as per (McGuire et al., 2023), can be summarized as:

$$\text{Sustainability(SSE)} = \{ \text{Process\_Sustainability}, \text{Product\_Sustainability} \}$$

where

$$\text{Product\_Sustainability} = f(\text{Environmental}(strata), \text{Social}(strata), \text{Economic}(strata), \text{Technical}(strata))$$

2. Key Technical and Non-Technical Principles

SSE is both a software quality attribute and an organizational capability, incorporating:

• Non-functional requirements: Energy consumption, resource optimization, durability ("perdurability": modifiability, adaptability, reusability) (Calero et al., 2013).
• Best practices in software engineering: Modular, testable, readable, standardized, and openly available code (Souza et al., 2019).
• Socio-organizational conditions: Collaborative development, ongoing training, institutional policy, and community support structures (Cohen et al., 2018).

Technical and non-technical factors are equally weighted, as ignoring either threatens long-term sustainability (Imran et al., 2019). Sustainability must be embedded at all SDLC stages and reflected in processes such as requirements engineering, version control, user experience, leadership, and funding.

3. Measurement, Evaluation, and Standardization

SSE emphasizes systematic, quantifiable evaluation frameworks:

• Quality benchmarks: The ISO/IEC 25010 model, extended to include sustainability sub-characteristics (e.g., energy consumption, resource optimization, perdurability) (Calero et al., 2013).
• Empirical metrics and benchmarking: Measurement of energy consumption, carbon footprint, and efficiency must be standardized and reproducible. Techniques range from coarse (system-wide Joule estimates) to fine-grained (per-operation profiling). Amortized energy metrics, such as $$E_{\text{amortized}} = E_{\text{training}} / N_{\text{inferences}}$$, facilitate balanced assessment of system sustainability (Cruz et al., 2 Jun 2025).
• Taxonomies of sustainability weaknesses: Machine-actionable knowledge bases catalog code patterns and design decisions directly detrimental to energy efficiency; careful, case-by-case analysis is required to tag these accurately (Pathania et al., 10 Jun 2025).

Standardization across tools, methodologies, and reporting formats (including hardware normalization and structured knowledge bases) is seen as crucial for both research and industrial practice (Cruz et al., 2 Jun 2025, Pathania et al., 10 Jun 2025).

4. Lifecycle Integration and Methodological Approaches

SSE techniques are positioned throughout the software lifecycle, including:

• Requirements engineering: The SDG Assessment for Requirements Elicitation (SDGARE) process translates United Nations SDG targets directly into actionable requirements. This approach merges traditional business-as-usual elicitation with top-down sustainability goals using ontologically-structured methods (Brooks, 2020).
• Agile and DevOps integration: Sustainability can be explicitly incorporated into agile backlogs, user stories, acceptance criteria, and regular retrospectives (Oyedeji et al., 9 Jul 2024). DevOps-oriented frameworks, such as SusDevOps, make sustainability a first-class principle, embedding environmental and social KPIs in monitoring, goal alignment, prioritization, and planning (David, 2023, Herati et al., 11 Mar 2025).
• Architectural patterns: Design for runtime adaptation, modularity, and traceability ensures sustainability is maintained during dynamic operations, particularly for AI-enabled and data-intensive systems (Cruz et al., 2 Jun 2025).
• Education and training: Teaching strategies combine classroom instruction, project-based learning, and experiential activities focused on robust software design, test-driven development, and community practices to support sustainability as habitual behavior (Gallagher et al., 7 Feb 2024, Penzenstadler et al., 2018).

5. Organizational Structures, Policy, and Community

SSE adoption depends on multidimensional support:

• Pillars of sustainability in research and practice: Software engineering, community, training, and policy underpin long-term sustainability, especially in research software engineering environments (Cohen et al., 2018).
• Context-sensitive interventions: Organizational context, particularly sectoral needs (e.g., financial services), shape how sustainability is perceived and operationalized. Sustainable practices must be co-designed with both management and practitioners to reconcile technical-economic drivers with human-centric concerns, such as workload and team health (Ghanta et al., 15 Sep 2025).
• Dedicated roles and teams: Some organizations prefer dedicated sustainability teams analogous to security or compliance units, but there is a risk of delegating responsibility rather than integrating sustainability as a shared value (Ghanta et al., 15 Sep 2025).
• Holistic organizational assessment: Comprehensive models, such as the Sustainability Index $S = \alpha H + \beta E + \gamma En$, aggregate human, economic, and environmental subscores, with tunable weights to match organizational priorities (Calero et al., 11 Jun 2024).

6. Challenges, Open Issues, and Research Directions

SSE faces several persistent challenges:

• Definitional ambiguity: There is neither a universally agreed ontology nor consensus on what constitutes "sustainable software," complicating measurement and cross-contextual comparison (Calero et al., 11 Jun 2024).
• Lack of empiricism and scaling: Much of the empirical literature remains nascent or absent, especially regarding process-level interventions, the social or human impact of SSE, and controlled evaluation of sustainability outcomes (McGuire et al., 2023).
• Integrating conflicting objectives: Trade-offs between technical performance (e.g., accuracy vs. energy use), short-term profitability vs. long-term sustainability, and developer well-being vs. rapid delivery require deliberative prioritization and transparent KPIs (David, 2023, Heldal et al., 2023).
• Education and skills gaps: Sustainability is underrepresented in higher education curricula, with lack of mature teaching materials and systematic integration cited as major barriers (Torre et al., 2017).

Key research fronts include: standardized measurement and reporting, empirical methodologies for process evaluation, tools for continuous sustainability monitoring, dedicated educational curricula, and the development of shared ontologies bridging technical, economic, and human factors (Cruz et al., 2 Jun 2025, Calero et al., 11 Jun 2024, Ghanta et al., 15 Sep 2025).

7. Sectoral Case Studies and Practical Recommendations

SSE practices must adapt to context:

• Domain-driven adaptation: In financial services, sustainability is shaped by regulatory requirements, high-availability needs, and data-driven operations, with management emphasizing cloud migration and developers highlighting workload management (Ghanta et al., 15 Sep 2025).
• Developing regions: Unique infrastructural challenges, such as unreliable electricity and reliance on fossil-fueled generators, frame sustainability as both an economic and environmental concern, which must be addressed through localized frameworks and practitioner education (Adisa et al., 2022).
• AI-enabled systems: Sustainability research agendas prioritize energy assessment standardization, hardware-normalized benchmarking, and life-cycle–spanning adaptation, with calls for longitudinal empirical practice and system-level integration (Cruz et al., 2 Jun 2025).

Practitioner recommendations across sectors include context-sensitive, co-designed interventions, integrating sustainability metrics into daily processes, and transparent communication aligning top-level goals with ground-level realities (Ghanta et al., 15 Sep 2025, Calero et al., 11 Jun 2024).

---

Sustainable Software Engineering is evolving from abstract principles and qualitative concerns toward measurable, standardized, and context-sensitive methodologies that integrate sustainability across technical, human, organizational, and societal strata. Empirical research, harmonized metrics, lifelong education, and broad interdisciplinary collaboration are critical for enabling software to fulfill both present and future responsibilities within a rapidly changing technological and environmental landscape.