Application-Aware Model for Peer Selection - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Rapport (Rapport De Recherche) Année : 2006

Application-Aware Model for Peer Selection

Résumé

We introduce in this paper the notion of application-aware optimal peer selection. With the advent of P2P and overlay networks, many applications of our days need to select the best peer to contact, either to transfer some data or to be positioned in an overlay network. This selection is still with no clear solution given the heterogeneity of the Internet in terms of path characteristics and access link speed, and the diversity of application requirements. Most of existing protocols rely on simple heuristics as for example choosing the closest peer in terms of delay. We believe that the selection of the best peer should be a function of more parameters (delay, bandwidth, loss rate) and subject to some application utility function (e.g., delay vs. bandwidth). This work aims at motivating the need for taking application requirements coupled with network path measurements into account during the peer selection process. The work consists of running extensive measurements over the PlanetLab overlay network and comparing different peer selection policies. One observation made in this work is that over PlanetLab, the best peer to contact is not always the closest one and it changes with the application needs. Another observation is that path characteristics are not strongly correlated with each other (e.g., smaller delay does not always mean larger bandwidth) which makes their separate use insufficient.
Fichier principal
Vignette du fichier
RR-5587.pdf (181.17 Ko) Télécharger le fichier
Loading...

Dates et versions

inria-00070420 , version 1 (19-05-2006)

Identifiants

  • HAL Id : inria-00070420 , version 1

Citer

Mohammad Malli, Chadi Barakat. Application-Aware Model for Peer Selection. [Research Report] RR-5587, INRIA. 2006, pp.19. ⟨inria-00070420⟩
82 Consultations
52 Téléchargements

Partager

Gmail Facebook X LinkedIn More