Monitoração de tráfego par-a-par em tempo real

The traffic due to P2P traffic-swapping applications has grown considerably in the last years and, although it is responsible for the biggest portion of Internet traffic nowadays, there are not many tools that aid monitoring and thus understanding the said traffic, both from an academic and from a p...

Descripción completa

Detalles Bibliográficos
Autor: Tiago Alves Macambira
Tipo de recurso: tesis de maestría
Estado:Versión publicada
Fecha de publicación:2005
País:Brasil
Institución:Universidade Federal de Minas Gerais (UFMG)
Repositorio:Repositório Institucional da UFMG
Idioma:portugués
OAI Identifier:oai:repositorio.ufmg.br:1843/RVMR-6EAJ3Z
Acceso en línea:http://hdl.handle.net/1843/RVMR-6EAJ3Z
Access Level:acceso abierto
Palabra clave:redes par-a-par
Computação móvel
Computação
Cliente/servidor (Computação)
Engenharia de tráfego
Redes de computação
Descripción
Sumario:The traffic due to P2P traffic-swapping applications has grown considerably in the last years and, although it is responsible for the biggest portion of Internet traffic nowadays, there are not many tools that aid monitoring and thus understanding the said traffic, both from an academic and from a practical perspective. In this work we resent Palant´ýr, a system for real-time stateful monitoring of P2P traffic in high-speed networks, and discuss the requirements and challenges faced during its construction. Even at 500 Mbps, a traffic rate higher than those commonly found at environments where Palant´ýr was supposed to be employed, it archives a packet drop rate of approximately 3.3 %and performs 99% better than what is possible with commonly used approaches. Even at those extreme load conditions Palant´ýr is capable of reassembling as much as 70 % of the whole data available in our experiments, a 100 fold gain of what is possible with traditionalsystems. We also present a case study where we analysed and characterized, with the aid of Palant ´ýr, the traffic of two P2P networks in a local Broadband ISP during a 10 days period. The results of this characterization suggest that there is a significant difference in the nature of both the resources traded by the users of the two P2P networks and of their traffic.