Green in software engineering: tools, methods and practices for reducing the environmental impacts of software use : a literature review

Background: The production, operation and use of information technology (IT) have a significant impact on the environment. As an example, the estimated footprint of global greenhouse gas emissions of the IT industry, including the production, operation and maintenance of main consumer devices, data centres and communication networks, doubled between 2007 (1–1.6%) and 2016 (2.5–3.1%). The European Union regulates the energy efficiency of data centre hardware. However, there is still a lack of regulation and guidance regarding the environmental impacts of software use, i.e. impacts from the production, operation and disposal of hardware devices required for using software. Aims: The goal of this thesis is to provide actionable knowledge which could be used by software practitioners aiming to reduce the environmental impacts of software use. Method: We conducted a systematic literature review of academic literature where we assessed evidence of the effectiveness of tools, methods and practices for reducing the environmental impacts of software use. The review covers 20 papers. Results: 60% of studied papers focus on reducing the energy consumption of software that is executed on a single local hardware device, which excludes networked software. The results contain 6 tools, 25 methods and 11 practices. Program code optimisation can potentially reduce the energy consumption of software use by 2–62%. Shifting the execution time of time-flexible data centre workloads towards times when the electric grid has plenty of renewable electricity can potentially reduce data centre CO2 emissions by 33.7%. Conclusions: The results suggest that the energy consumption of software use has received much attention in research. We suggest more research to be done on environmental impacts other than energy consumption, such as CO2 emissions, software-induced hardware obsolescence, electronic waste and freshwater consumption. Practitioners should also take into account the potential impacts of data transmission networks and remote hardware, such as data centres, in addition to local hardware