Empirical characterization of Internet Exchange Points
Today's public Internet eXchange Points (IXP) are an important factor in the Internet ecosystem, nevertheless, despite their importance, the community still lacks empirical data on the nature of the traffic exchanged through IXPs. We strive in obtaining a deeper understanding of exchange points...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Fecha de publicación: | 2011 |
| País: | España |
| Institución: | IMDEA Networks Institute |
| Repositorio: | IMDEA Networks Institute Digital Repository |
| Idioma: | inglés |
| OAI Identifier: | oai:dspace.networks.imdea.org:20.500.12761/1034 |
| Acceso en línea: | http://hdl.handle.net/20.500.12761/1034 |
| Access Level: | acceso abierto |
| Palabra clave: | Q Science::Q Science (General) Q Science::QA Mathematics::QA75 Electronic computers. Computer science T Technology::T Technology (General) T Technology::TA Engineering (General). Civil engineering (General) T Technology::TK Electrical engineering. Electronics Nuclear engineering |
| Sumario: | Today's public Internet eXchange Points (IXP) are an important factor in the Internet ecosystem, nevertheless, despite their importance, the community still lacks empirical data on the nature of the traffic exchanged through IXPs. We strive in obtaining a deeper understanding of exchange points by closely examining the history of a typical regional European IXP: the Slovakian-IX. By crawling the web archive we obtained several snapshots of the Slovak-IXP website from the last 14 years. This historical data allows us to study the dynamics of the IXP in terms of: the (low-tier) AS-level topology, the traffic dynamics, the port capacity and the traffic matrix. Our datasets show that, different from Customer-Provider links, peering links are very stable as we observe that once they are created they are very unlikely to disappear. After the proliferation of content service providers in 2006, the traffic profiles of the ISPs peering at the IXP experienced significant changes reflected in the extreme growth of the content ISPs traffic and heavily imbalanced traffic ratios. An analysis on the distribution of traffic of the IXP hints that since the beginning of SIX a small fraction of heaviest peering pairs carry the majority of the SIX traffic while most of the peering pairs exchange very low amount of traffic and hence enjoy close-to-zero monetary gain. We strongly believe that in contrast with confidential datasets typically used in studying the Internet traffic characteristics, IXP's data provide rich and publicly available resources crucial for understanding various aspects of the Internet. |
|---|