Análise da distribuição de streaming media em arquiteturas do tipo peer-to-peer

• Ano de defesa: 2003
• Autor(a) principal: Marisa Affonso Vasconcelos
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Ciência da computação; Recuperação de dados (Computação); Computação; Cliente/servidor (Computação); Redes locais de computação
• Link de acesso: http://hdl.handle.net/1843/BUOS-9KRN8P

Client/server architectures have shown to be inefficient in distributing streaming media. The maximum number of users is limited due to the computing resources, which are more demanding than in traditional web applications. Onde of the proposals for an efficient distribution is the use of a Peer-to-Peer networks (P2P). In this proposal, the content distribution is made by the clients themselves, which are called servents, since they act as client and server at the same time. Most of the studies in the literature evaluate this scenario through simulation, so there is a lack of quantitative analysis of the requirements for server and network resources in actual P2P systems. The goal of this work is to fill this gap by providing experimental results that quantify the savings in server andnetwork resources if a P2P approach is adoted. Towards this goal, we have built an experimental testbed to evaluate both architectures, client/server e P2P, with varying the number of clients during the experiments. This experimental testbed is composed of one new efficient and scalable application called streaming servent, one streaming serverapplication (Darwin) and a workload consisting three media files. In this work, we use simple analytical formulas to evaluate the scalability of our servent application. The results show that the use of a simple two-level architecture reduced the load on the server by more than 60%, providing better stream quality and scalability.