Detalhes bibliográficos
Ano de defesa: |
2014 |
Autor(a) principal: |
Cruz, Mário Augusto da
 |
Orientador(a): |
Cardoso, Kleber Vieira
 |
Banca de defesa: |
Cardoso, Kleber Vieira,
Corrêa, Sand Luz,
Rosa, Thierson Couto,
Abelém, Antônio Jorge Gomes |
Tipo de documento: |
Dissertação
|
Tipo de acesso: |
Acesso aberto |
Idioma: |
por |
Instituição de defesa: |
Universidade Federal de Goiás
|
Programa de Pós-Graduação: |
Programa de Pós-graduação em Ciência da Computação (INF)
|
Departamento: |
Instituto de Informática - INF (RG)
|
País: |
Brasil
|
Palavras-chave em Português: |
|
Palavras-chave em Inglês: |
|
Área do conhecimento CNPq: |
|
Link de acesso: |
http://repositorio.bc.ufg.br/tede/handle/tede/4372
|
Resumo: |
Software Defined Networking represents a new paradigm that eases the operation, monitoring and network managing through the decoupling between the control plane and the data plane. However, in this new context, some classic solutions in the network monitoring field need to be revisited, as there are new constraints, but there are also new opportunities. In monitoring context, one strategy commonly used, mainly in high capacity networks, is the tracking of the most frequent items, also known as heavy hitters. One approach to monitoring the most frequent items consists in detecting the hierarchical heavy hitters, which allows an efficient real time monitoring. In this work, we propose and evaluate a new monitoring solution capable of online detection of hierarchical heavy hitters, using the characteristics of software defined networks, in special the OpenFlow protocol. Our proposal, combines a flexible accounting of flow rules, from OpenFlow switches, with inspection of traffic samples through a dedicated device. We evaluate our proposal in a simulated and emulated environments, both using packet traces generated artificially and also from real networks. The results show that our proposal has satisfactory accuracy and low convergence time in comparison to a previous solution to OpenFlow networks, in addition to identify heavy hitters in two dimensions. |