Engenharia de tráfego para obtenção de QoS na comunicação entre tarefas em grades computacionais

• Ano de defesa: 2006
• Autor(a) principal: Torres, Guilherme Mundim
• Orientador(a): Guardia, Hélio Crestana
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação - PPGCC
• País: BR
• Palavras-chave em Português: Redes de comunicação de dados; Engenharia de tráfego; Qualidade de serviço; Grade computacional; Multi-protocol label switching
• Palavras-chave em Inglês: Traffic engineering; MPLS; QoS; Grids
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/364

The advent of grid computing made possible to access distributed resources, even when they are geographically spread or belong to different organizations. The most used environment for accessing these distributed resources is the Internet, a worldwide computer network based in TCP/IP architecture. Grid computing provides the infrastructure necessary for managing and communicating with the resources offered by different organizations. These organizations are also known as virtual organizations (VO's). Some of the applications used in these collaborating environments may have minimum requirements by quality of service (QoS). However, the "best effort" service, which is offered by Internet, is not capable to satisfy these QoS requirements. In this case, a different solution is needed, in order to provide guarantees related to the traffic in communication channels. This master thesis aims to apply the concepts of quality of service for networks in grid computing, providing end-to-end quality of service between grid computing applications. In order to achieve this goal, we investigate the use of commutation infrastructure provided by MPLS networks. Using traffic engineering mechanisms for routes determination, we aim to provide better control of data flows, improving the performance of distributed applications in geographically highly spread environments.