Bootstrap-Based Estimation of Flow-Level Network Traffic Statistics

Stenio FernandesTatiene SouzaCarlos KamienskiDjamel H. Sadok

Network traffic measurement and analysis have been playing an important role in traffic engineering, and network planning. Most recent research papers in this field have been focusing to passive flow measurement since collecting packet-level data in high-speed links is prohibitively expensive. Although there are techniques for handling flow statistics in modern routers, transmitting and storing flow-level information still imposes a significant burden on the network management operation. In this paper, we advocate that only a small portion of the flow records need to be preserved for further processing. We propose the use of the Bootstrap resampling technique for deriving properties from a previously pre-processed sampled set of flows. Our results show that only 10% or less of the original sampled statistics is necessary in order for Bootstrap to reconstruct the characteristics of the original raw flow records. We also show that such technique has a low computational overhead time.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: