A passive available bandwidth estimation methodology
The Available Bandwidth (AB) of an end-to-end path is its remaining capacity and it is an important metric for several applications such as overlay routing and P2P networking. That is why many AB estimation tools have been published recently. Most of these tools use the Probe Rate Model, which requi...
| Autores: | , , , |
|---|---|
| Formato: | informe técnico |
| Fecha de publicación: | 2009 |
| País: | España |
| Recursos: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/14354 |
| Acesso em linha: | https://hdl.handle.net/2117/14354 |
| Access Level: | acceso abierto |
| Palavra-chave: | Telecommunication -- Traffic Passive Measurement Available bandwidth Telecomunicació -- Tràfic Àrees temàtiques de la UPC::Enginyeria de la telecomunicació::Telemàtica i xarxes d'ordinadors |
| Resumo: | The Available Bandwidth (AB) of an end-to-end path is its remaining capacity and it is an important metric for several applications such as overlay routing and P2P networking. That is why many AB estimation tools have been published recently. Most of these tools use the Probe Rate Model, which requires sending packet trains at a rate matching the AB. Its main issue is that it congests the path under measurement. We present a different approach: a novel passive methodology to estimate the AB that does not introduce probe traffic. Our methodology, intended to be applied between two separate nodes, estimates the path’s AB by analyzing specific parameters of the traffic exchanged. The main challenge is that we cannot rely on any given rate of this traffic. Therefore we rely on a different model, the Utilization Model. In this paper we present our passive methodology and a tool (PKBest) based on it. We evaluate its applicability and accuracy using public NLANR data traces. Our results -more than 300Gb- show that our tool is more accurate than pathChirp, a state-of-the-art active PRM-based tool. At the best of the authors’ knowledge this is the first passive AB estimation methodology. |
|---|