Modélisation conséquentielle de la consommation d'énergie d'un groupe de serveurs générant un nuage informatique et attributionnelle des bénéfices de la virtualisation
Year: 2014
Type: dissertation
Abstract: Nowadays, electricity is an essential element of our modern society, people use it every day in their daily activities. However, its production is responsible for numerous environmental damages accounting for 30% of the global greenhouse gas (GHG) emissions, recognized by the majority of the scientific community to be the origin of the climatic changes. Data center can be considered as supercomputers and represent already 1,3 % of the global electricity consumption. It is a very high portion considering the youth of this technology. To decrease their consumption, different strategies of optimization exist, for example the virtualization, which makes it possible to install several operating systems known as virtual machines (VM) on a single server, increasing its utilization. But virtualization has a cost due to the utilisation of a programme, named “hypervisor”, which connects the VM to the hardware. To choose adequately the strategy to apply, it is necessary to possess an environmental evaluation tool, allowing the quantification of the environmental gains and costs of each optimization approaches. Life cycle assessment (LCA) seems to satisfy these conditions because it is a methodology that quantifies the potential environmental impacts of the entire life cycle of a product or service. However, LCA is a methodological tool still in development and many issues are in discussion. This master thesis aims to quantify the environmental impacts induced by the installation of data centers in Canada and to evaluate the environmental benefits of using virtualization. The problematic is divided in two parts: first, how to model the environmental impacts of new data centers installed in Canada, and second, how to quantify the benefits of virtualization? Two objectives are fixed to solve the two parts of the problematic: evaluate, by combining consequential LCA with an economic model, the environmental impacts of energy consumption of data centers implanted in Canada and evaluate with a LCA approach the environmental benefits of virtualization. The first part of the methodology develops a consequential and prospective approach, combining LCA with the economic model E3MC (which describe the North American energy sector and the Canadian economy), to evaluate the impact on the North American energy sector of datacenters implantation in Canada. A LCA is prospective when a future system is evaluated and it is consequential when impacts from a change due to a decision are evaluated. This analysis is
Author: Nathan Vandromme
Language: French
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W400273441