Software sustainability assessing the environmental impact of the software life cycle
With the rapid rise of laptops, smartphones, and artificial intelligence, there has been limited attention paid to their environmental impact. This thesis presents a framework developed after extensive research into the green software domain, aimed at helping corporations with emissions accounting a...
| Author: | |
|---|---|
| Format: | master thesis |
| Publication Date: | 2024 |
| Country: | España |
| Institution: | Universitat Politècnica de Catalunya (UPC) |
| Repository: | UPCommons. Portal del coneixement obert de la UPC |
| Language: | English |
| OAI Identifier: | oai:upcommons.upc.edu:2117/419400 |
| Online Access: | https://hdl.handle.net/2117/419400 |
| Access Level: | Open access |
| Keyword: | Sustainability Computer software-- Development Software, Green Software, Sustainability, IT, CSRD, Corporate Sustainability, Software Engineering, Software Sustainability Sostenibilitat Programari--Desenvolupament Àrees temàtiques de la UPC::Desenvolupament humà i sostenible |
| Summary: | With the rapid rise of laptops, smartphones, and artificial intelligence, there has been limited attention paid to their environmental impact. This thesis presents a framework developed after extensive research into the green software domain, aimed at helping corporations with emissions accounting and reporting, while adhering to the Corporate Sustainability Reporting Directive (CSRD) and similar policies. The core metric adopted within the framework is Software Carbon Intensity (SCI), which is measured through system utilization parameters. These parameters are calibrated using two benchmarking tools, Running Average Power Limit (RAPL) and Performance Counter Monitor (PCM). Two case studies were conducted to demonstrate the framework’s practical application. In the first case, representing the software development stage, emissions of approximately 330 gCO2eq were generated during typical office hours from 09:00 to 18:30, which extrapolates to 9.87 kgCO2eq per developer per sprint. The second case study, focused on software usage, recorded emissions ranging from 0.054 gCO2eq to 2.73 gCO2eq per run, depending on the location of execution and the method of carbon intensity data aggregation. A reporting method using candlestick charts is also introduced, providing companies a useful tool for Scope emissions accounting. |
|---|