Sustainability Awareness Framework
- Sustainability Awareness Framework is a structured approach that makes sustainability concerns visible by integrating social, environmental, economic, technical, and individual dimensions.
- It operationalizes sustainability by converting broad concerns into measurable indicators, metrics, and actionable design principles across domains like software engineering and manufacturing.
- The framework fosters awareness formation and behavior change through targeted pedagogical strategies, real-world case studies, and stakeholder engagement.
Sustainability Awareness Framework denotes a structured approach for making sustainability concerns visible in analysis, design, assessment, and education. In a Software Engineering curriculum case study, the term refers to a framework used with “the five dimensions of the Sustainability Awareness Framework,” targeted discussion questions, and good practice examples from Green Software Foundation patterns to identify where sustainability can be embedded across modules (Chitchyan et al., 1 Oct 2025). Across adjacent literature, related frameworks serve broader purposes: they model sustainability as a systemic and multidimensional concern, translate goals into indicators and data requirements, connect sustainability to requirements and architecture decisions, and support observability, pedagogy, and behavior change (Becker et al., 2014, Lago et al., 2024).
1. Conceptual foundations
A central foundation is the claim that sustainability is not a narrow environmental add-on but a systemic property. The Karlskrona Manifesto states that sustainability “has to be understood on a set of dimensions, including social, individual, environmental, economic, and technical,” and that design must distinguish “Direct, first order effects,” “Indirect, second order effects,” and “Systemic, third order effects” (Becker et al., 2014). The same source rejects several recurrent misconceptions: sustainability is not only about protecting the environment or maintaining business activity, is not a separate specialty with a few defined connections to software, is not reducible to immediate effects, and is not a late-stage tradeoff against a fixed set of quality attributes (Becker et al., 2014).
In software architecture research, the Sustainability Assessment Framework Toolkit adopts a compatible view by treating sustainability as a software quality property and defining it as preserving intended benefits over time. Its core model distinguishes a sustainability focus dimension—technical, economic, social, and environmental—from a time dimension of first-order, second-order, and third-order effects (Lago et al., 2024). The toolkit also distinguishes sustainability awareness from sustainability as a property: awareness is the capacity to notice, discuss, model, and reason about sustainability impacts during requirements and architecture work, whereas sustainability itself concerns long-term preservation of beneficial use (Lago et al., 2024).
Taken together, these works suggest that a Sustainability Awareness Framework is less a single canonical artifact than a family of structures that make sustainability discussable, modelable, and actionable across domains. What remains stable is the insistence on multidimensionality, long time horizons, and visibility of both the system itself and the wider context in which it operates (Becker et al., 2014, Lago et al., 2024).
2. Structural forms and recurring dimensions
Frameworks labeled or functioning as sustainability-awareness structures differ in how they decompose the problem. Some are principle-driven, some are indicator-driven, and some are explicitly architectural or workflow-oriented.
| Framework | Core structuring elements | Domain |
|---|---|---|
| Karlskrona Manifesto | five dimensions; first-, second-, third-order effects | software design |
| SAF Toolkit | Decision Map; Sustainability-Quality Model | software architecture |
| S-Assessment Tool | five sections plus profiling section | organizational social sustainability |
| USFM | manufacturing process and system; data process; KPI selection and assessment | manufacturing |
| Observability blueprint | collection; aggregation and storage; processing; visualization | software systems |
The Karlskrona Manifesto defines five dimensions—individual, social, environmental, economic, and technical—and couples them to effect orders and system visibility (Becker et al., 2014). The SAF Toolkit uses four dimensions—technical, economic, social, environmental—and organizes its instruments around a Decision Map and a Sustainability-Quality Model with quality attributes, metrics, and dependency matrices (Lago et al., 2024). SEER also uses four dimensions—environmental, economic, social, and technical—but operationalizes them as a general sustainability requirements taxonomy with the fields Sustainability Requirement, Dimension, and Category (Roy et al., 10 Oct 2025).
Other frameworks are explicitly domain-shaped. The S-Assessment Tool for Social Sustainability is organized into Health and Wellness, Gender Equality, Decent Work and Economic Growth, Reducing Inequalities, and Responsible Production and Consumption (Annarelli et al., 2024). The Unified Smart Factory Model integrates Manufacturing Process and System, Data Process, and KPI Selection and Assessment in a single framework (Kaushal et al., 11 Dec 2025). FAIR+S extends FAIR and FAIR4RS with a fifth dimension, “S = Sustainable,” expressed through principles S1–S5 concerning energy attributes, benchmarks, alignment with sustainability frameworks, transparency and accountability, and life-cycle sustainability (Valko et al., 17 Jun 2026).
The curriculum case study centered on Software Engineering reports a five-dimensional Sustainability Awareness Framework, but the abstract does not enumerate the dimension names (Chitchyan et al., 1 Oct 2025). That omission is significant because it shows that the label “Sustainability Awareness Framework” can denote a recognizable pedagogical scaffold even when only its role, rather than its internal taxonomy, is disclosed.
3. Operationalization, metrics, and assessment logic
Awareness frameworks become operational when they convert broad sustainability concerns into indicators, scores, or explicit dependency structures. A systematic review of organizational sustainability assessment frameworks reports that indicator selection commonly relies on literature reviews, stakeholder interviews, questionnaires and surveys, Pareto analysis, SMART, Delphi and fuzzy-Delphi methods, materiality analysis, and alignment with standards such as GRI and ISO. For weighting, fuzzy-AHP is highlighted as robust; for interdependency analysis, DEMATEL, VIKOR, correlation analysis, and causal models are common, but their treatment of sustainability dynamics is described as static (Farahdel et al., 2024).
The S-Assessment Tool is a concrete example of operationalization. It converts social sustainability themes into a questionnaire whose section scores are “calculated as the sum of the individual points attributed to each answer,” compared to the maximum achievable score in each section, and then mapped to four organizational categories: Pioneer for scores less than or equal to 30%, Builder for 31%–60%, Transformer for 61%–80%, and Leader for more than 81% (Annarelli et al., 2024). The framework is explicitly aligned with the UN 2030 Agenda and the SDGs, but its current validation is described as benchmark testing rather than a fully reported psychometric study (Annarelli et al., 2024).
Pandemic policy analysis offers a different assessment logic. Its yearly framework uses 20 human activity/impact parameters grouped into Negative, Limiting, and Positive classes, displayed as an infographic bar chart. Negative and positive parameters are normalized to the range $0$ to $1$, while limiting parameters are normalized to to $1$. Negative values should move toward zero, positive values should increase, and limiting variables are treated as policy levers for region-specific upstream and downstream intervention (Banerjee et al., 2020).
At the most formal end, STRIDE defines benchmark trust for sustainability-rating datasets through the latent score
where is credibility, reliability, intimacy, and self-served purpose (Cai et al., 19 Feb 2026). This formulation makes explicit that awareness frameworks can be used not only to measure sustainability outcomes, but also to evaluate the trustworthiness of the datasets and procedures used to judge them.
4. Integration across design, requirements, runtime, and life cycle
A defining feature of recent frameworks is movement toward earlier and more continuous integration. In materials design, sustainability-informed materials design argues that life cycle thinking should enter “at the earliest stages of materials development,” and recasts the ISO life cycle assessment phases—goal and scope, life cycle inventory, life cycle impact assessment, and interpretation—for low-TRL discovery under uncertainty. Its governing principles are bottom-up modeling, treating uncertainty as information, iteration across TRLs, decision relevance rather than numerical precision, and interoperable modular data and tools (Woods-Robinson et al., 24 Apr 2026).
In software engineering, SEER places sustainability directly in requirements engineering. Its workflow has three stages: identification of sustainability requirements from a general taxonomy, evaluation of how existing system requirements relate to those sustainability requirements, and optimization of requirements that fail to satisfy them (Roy et al., 10 Oct 2025). The framework uses agentic RAG, knowledge-graph retrieval from standards, a fine-tuned sentence-transformer model for related-pair extraction, and multi-agent LLM reasoning for relation classification and requirement revision (Roy et al., 10 Oct 2025).
Architecture-level integration is represented by the SAF Toolkit and by observability-oriented work. The SAF Toolkit uses a Decision Map to capture sustainability-related design concerns and an SQ Model to map quality attributes to sustainability dimensions and metrics, including inter- and intra-dimensional dependencies (Lago et al., 2024). The architectural blueprint for sustainability observability structures deployment as a layered pipe-and-filters system with collection, aggregation and storage, processing, and visualization. Its processing loop is described as “gather, store, distill, and clear,” and its special case of energy measurement shows how sustainability awareness can be turned into operational telemetry (Toczé et al., 10 Apr 2026).
Manufacturing and distributed AI research extend the same logic. USFM translates high-level sustainability goals into factory-level indicators through a top-down sequence from organizational goals and KPI selection to OPM-based factory modeling, metric establishment, data collection, analysis, and control (Kaushal et al., 11 Dec 2025). GreenDFL quantifies energy and carbon across the decentralized federated learning lifecycle through
with separate accounting for training, communication, and aggregation, and then uses sustainability-aware aggregation and node-selection algorithms to reduce footprint (Feng et al., 27 Feb 2025). FAIR+S, finally, embeds sustainability metadata into research data and software stewardship rather than treating environmental accountability as a separate reporting silo (Valko et al., 17 Jun 2026).
5. Awareness formation, pedagogy, and behavior change
Several frameworks explicitly address how awareness is formed rather than merely measured. The Software Engineering curriculum case study reports a collaborative and iterative consultative process in which academics and students co-identified module-level opportunities for embedding sustainability, guided by the Sustainability Awareness Framework, targeted discussion questions, illustrative examples, and Green Software Foundation patterns (Chitchyan et al., 1 Oct 2025). The emphasis is on distributed integration across modules rather than a single standalone sustainability topic (Chitchyan et al., 1 Oct 2025).
Teaching-oriented work in mathematics follows a similar pattern. One course narrative frames sustainability awareness through “interactive problem-solving sessions, group discussions, and short projects,” beginning from examples of unsustainability and moving toward mathematical modeling of climate, ecosystems, pollution, energy, and tipping points (Shahgholian, 14 Oct 2025). The pedagogical principle that “Sustainability is best defined by its absence” is paired with the more formal statement that sustainability “refers to the ability to maintain or support a process continuously over time” (Shahgholian, 14 Oct 2025).
The Cognitive-Affective-Systemic framework extends awareness beyond discursive and quantitative forms. It defines sustainability awareness as the interaction of cognitive comprehension, affective resonance, and systemic understanding, with “logomotion” naming the state in which understanding and feeling move together (Liu, 20 Oct 2025). EcoEcho applies a different mechanism: players perform unsustainable actions in a GenAI-mediated game world and then immediately see environmental degradation. In a study with 23 participants, intended ecological behavior increased significantly, while sustainability attitudes improved only slightly and not significantly (Zhang et al., 2024).
Behavioral theory supplies an additional formal pathway. A TPB-based conceptual framework on sustainable entrepreneurship proposes the sequence Climate change campaigns $1$0 climate awareness $1$1 sustainable development values $1$2 attitude, subjective norms, and perceived behavioral control $1$3 sustainable entrepreneurial intention, with climate awareness also moderating the value-to-TPB links (Islam et al., 2024). This makes explicit that awareness alone is not treated as sufficient; it must be translated into values and then into intention.
Large-scale field analysis shows why such pedagogical and motivational mechanisms are considered necessary. In robotics research, mentions of sustainability-related impacts are typically below 2%, explicit SDG references remain below 0.1%, and the proportion of sustainability-motivated papers remains below 5%, even though a very large fraction of papers can be mapped to SDG-relevant domains (Skuric et al., 9 Apr 2026). The literature therefore identifies an awareness gap and a motivation gap rather than a lack of latent technical relevance (Skuric et al., 9 Apr 2026).
6. Misconceptions, limitations, and current directions
A recurrent controversy concerns what exactly awareness frameworks measure. The robotics survey is explicit that it measures communication and stated intent, not real-world impact (Skuric et al., 9 Apr 2026). EcoEcho reports significant change in intended sustainable behaviors but only slight movement in attitudes (Zhang et al., 2024). FAIR+S validation finds strong perceived importance but a gap between importance and feasibility, together with lower familiarity with green software practices than with FAIR or the SDGs (Valko et al., 17 Jun 2026). These results indicate that awareness, motivation, implementation capacity, and actual sustainability performance are related but non-identical constructs.
Another limitation concerns measurement overhead and methodological maturity. The observability blueprint discusses the “observer effect” qualitatively and cites prior work reporting that tracing can reduce throughput by up to 80% and increase latency by 175% in some scenarios (Toczé et al., 10 Apr 2026). The S-Assessment Tool does not report factor analysis, reliability coefficients, or formal construct-validity statistics in the excerpted account (Annarelli et al., 2024). STRIDE identifies its own unresolved issues: weights are not formally learned, thresholding rules are not calibrated, and some criteria remain discretized or simplified (Cai et al., 19 Feb 2026). SEER, similarly, depends on taxonomy coverage, product-scope quality, and LLM stability, and its evaluation is limited to four domains (Roy et al., 10 Oct 2025).
Current directions point toward more dynamic, hybrid, and institutionally embedded frameworks. Organizational review work recommends hybrid MCDM models, extension across organizational levels and supply chains, and periodic revision rather than static scorecards (Farahdel et al., 2024). The SAF Toolkit plans KPI models, alignment with ISO/IEC/IEEE 42010:2022, Green Lab integration, and dashboard support (Lago et al., 2024). FAIR+S emphasizes automation, minimal reporting requirements, training, and infrastructure support (Valko et al., 17 Jun 2026). In research infrastructure, the ErUM workshop report links awareness to measurement, teaching, governance, incentives, and “breathing” computing centers aligned with low-carbon energy availability (Bella et al., 27 Feb 2026). A position paper on sustainable AI makes a comparable argument at ecosystem scale: sustainable AI lies in the quadrant where climate awareness and resource awareness are both high, and reaching that state requires coordinated action from individual researchers to global governance (Bakhtiarifard et al., 27 Feb 2025).
In this sense, Sustainability Awareness Framework has evolved from a vocabulary for reflection into a design and governance pattern. Its mature forms do not merely ask whether sustainability matters; they specify which dimensions matter, when effects appear, how they are observed or scored, where intervention is possible, and which institutional conditions are required to move from awareness to sustained practice (Becker et al., 2014, Bella et al., 27 Feb 2026).