Navegar por los elementos (1 total)

  • Resumen es exacto "The present document introduces the research and implementation of different detection methods done on the capture platform of the Telecom Bretagne school of engineering in Brest, France. The methods developed are purely statistical based and as a premise, the payload of the packets can’t be used in order to perform the detection.
    As Internet becomes a social phenomenon, more services are using the network to work. ISPs and enterprises need a tool to classify traffic and in this way decide what to do with each flow of information. It will be seen how former and quite simple detection methods became obsolete, and the new different approaches to solve this problem.
    The developed methods will take different identifiers from the traffic captured and use different models to identify the kind of information flow, based on a prior learning process this platform must be put through. The learning process will result in different signatures which stand for each service.
    A Theoretical introduction to every concept used to develop the methods is included, as long as two large appendixes which describe how to configure, run and modify the application implemented.
    Finally, conclusions will be presented as well as the issues and considerations the ongoing research project should address in the future.
    The appendixes are the guides written to help the users and developers work with the developed application."

Título: Statistical methods for automatc Internet application recognition

Formatos de Salida

atom, csv, dc-rdf, dcmes-xml, json, omeka-xml, rss2