Masterarbeit MSTR-2024-15

Bibliograph.
Daten
Hoxha, Erisa: Energy and CO2 footprint of cloud/backend processing.
Universität Stuttgart, Fakultät Informatik, Elektrotechnik und Informationstechnik, Masterarbeit Nr. 15 (2024).
53 Seiten, englisch.
Kurzfassung

Data centers are the backbone of the digital infrastructure and they are responsible for storing and processing large amounts of data. In 2023, global emissions from cloud computing made up from 2.5% to 3.7% of worldwide GHG, thereby exceeding emissions from commercial flights (2.4%). The demand for data centers is increasing exponentially due to high number of users (5.8 billion in October 2023) and rising demand for services and technologies such as streaming, cloud gaming, blockchain, artificial intelligence and machine learning. The emission in this data centers is caused by electricity consumption (to run the servers), water consumption (to run the servers) and the lifetime of the equipment. Many companies have outsourced their tech infrastructure thus their information is being managed more efficiently, but this measure does not remove completely their environmental impact. Different measures are currently being taken from the top three cloud providers: AWS, Microsoft Azure and GCP. AWS leads the list with over 30% market share and they have committed to reaching net zero emissions by 2040. They announced that they will inform their customers about their carbon footprint while continuing to improve emissions release. Azure team is also working to improve their climate impact and are currently trying on a 2-year-old experiment where they have submerged 800 servers on the ocean floor of the Scottish coast. On the other side, GPC, which has the least market share among these three, claims to have already reached net zero emissions. Apart from there, there are many software that are using different algorithm to measure the carbon footprint for data centers.

Abteilung(en)Universität Stuttgart, Institut für Architektur von Anwendungssystemen, Architektur von Anwendungssystemen
BetreuerAiello, Prof. Marco
Eingabedatum2. Juli 2024
   Publ. Abteilung   Publ. Institut   Publ. Informatik