Submitted by Setor de Tratamento da Informa??o - BC/PUCRS (tede2@pucrs.br) on 2016-07-14T16:47:27Z
No. of bitstreams: 1
DIS_MARCELO_DA_SILVA_CONTERATO_COMPLETO.pdf: 3136466 bytes, checksum: acab99222ce3923954a284d165dc3f87 (MD5) / Made available in DSpace on 2016-07-14T16:47:27Z (GMT). No. of bitstreams: 1
DIS_MARCELO_DA_SILVA_CONTERATO_COMPLETO.pdf: 3136466 bytes, checksum: acab99222ce3923954a284d165dc3f87 (MD5)
Previous issue date: 2016-01-15 / Currently, Data Centers have their resources used at extremely high loads, which leads to the uncontrolled use of resources. It maintains a high energy consumption, even at times when traffic demand is low. At this time, researchers are performing several researches in energy efficiency for data center, however, most of the results were focused on two major components: servers and cooling systems. In this study, we propose strategies for data center network configuration to reduce energy consumption through the Software Defined Networking paradigm. Such strategies were combined with techniques to reduce energy consumption and evaluated compared to the power-agnostic environments. Besides, the impact of different fat-tree topology sizes and proposed strategies were simulated and compared taking into account the presented energy savings. By applying the overload factors of 1:5 and 1:20, energy-saving rate on the network reached 70.02% for a fat-tree topology size of k = 12, and 64.82% for a fat-tree topology size of k = 8, when compared to a traditional network. / Atualmente, Data Centers t?m seus recursos utilizados em cargas extremamente altas, o que leva ? utiliza??o de recursos de forma descontrolada. Isto mant?m um elevado consumo de energia, at? mesmo em momentos em que a demanda de tr?fego ? baixa. At? agora, os pesquisadores realizaram diversas pesquisas no campo da efici?ncia energ?tica para Data Centers, no entanto, a maior parte dos resultados se concentra em dois componentes principais: servidores e sistemas de refrigera??o. Neste trabalho, s?o propostas estrat?gias para configura??o da rede de Data Center visando reduzir o consumo de energia,
atrav?s do paradigma de SDN (Software Defined Networking). Tais estrat?gias foram combinadas com t?cnicas de redu??o do consumo de energia e avaliadas em compara??o ao consumo de energia em ambientes sem preocupa??o com economia de energia. Al?m disso, o impacto de diferentes tamanhos de topologias fat-tree e as estrat?gias propostas foram simuladas e comparadas quanto a economia de energia apresentada. Ao aplicarmos os fatores de sobrecarga de 1:5 e 1:20, a taxa de economia de energia na rede chegou a 70,02% com uma topologia fat-tree de tamanho k = 12, e a 64,82% com uma topologia fat-tree de k = 8, em compara??o com a rede tradicional.
| Identifer | oai:union.ndltd.org:IBICT/oai:tede2.pucrs.br:tede/6831 |
| Date | 15 January 2016 |
| Creators | Conterato, Marcelo da Silva |
| Contributors | Ferreto, Tiago Coelho |
| Publisher | Pontif?cia Universidade Cat?lica do Rio Grande do Sul, Programa de P?s-Gradua??o em Ci?ncia da Computa??o, PUCRS, Brasil, Faculdade de Inform?tica |
| Source Sets | IBICT Brazilian ETDs |
| Language | Portuguese |
| Detected Language | English |
| Type | info:eu-repo/semantics/publishedVersion, info:eu-repo/semantics/masterThesis |
| Format | application/pdf |
| Source | reponame:Biblioteca Digital de Teses e Dissertações da PUC_RS, instname:Pontifícia Universidade Católica do Rio Grande do Sul, instacron:PUC_RS |
| Rights | info:eu-repo/semantics/openAccess |
| Relation | 1974996533081274470, 600, 600, 600, -3008542510401149144, 3671711205811204509 |
Page generated in 0.0016 seconds