Skip to Main content Skip to Navigation

On the network awareness of P2P applications.

Abstract : This work originates in the context of the ''NapaWine'' European project and its main goal is to make P2P applications aware of the underlying network topology. This information (e.g., path capacity, latencies, etc.) helps to optimize P2P traffic and its benefits are twofolds: on one hand operators can limit traffic on peering links by confining traffic in their AS. On the other hand, users can experience higher quality of service due the proximity of neighbors. To tackle the problem we first analyze existent applications and we especially gauge their level of network awareness: since many systems are closed source, we study them as black boxes by means of purely passive analysis, then we setup a controlled testbed and finally we make use of parallel active probing to gather dynamic neighbors properties. As a second step we test state of the art chunk diffusion algorithms in a realistic simulator where we model latencies and access link capacities according to academic results. Within such a scenario we study how chunk diffusion performance suffers in presence of measurement errors and out-of-date system state knowledge. Finally we develop an emulation scenario in which we are able to test real applications. Our testbed can scale up to 200 application instances and emulates the topology of the existing research network Abilene; it also allows to perform traffic engineering which we exploit to analyze coupling phenomenon between IP and overlay routing.
Complete list of metadata

Cited literature [14 references]  Display  Hide  Download
Contributor : Paolo Veglia Connect in order to contact the contributor
Submitted on : Sunday, December 4, 2011 - 3:06:49 PM
Last modification on : Sunday, June 26, 2022 - 9:49:11 AM
Long-term archiving on: : Monday, December 5, 2016 - 7:52:22 AM


  • HAL Id : pastel-00647980, version 1


Paolo Veglia. On the network awareness of P2P applications.. Networking and Internet Architecture [cs.NI]. Télécom ParisTech, 2011. English. ⟨pastel-00647980⟩



Record views


Files downloads