An Experimental Study on Peer Selection in a P2P Network over PlanetLab

Abstract : Peer selection is an important aspect in many P2P applications requiring efficient assignment and execution of jobs to peer nodes and search and file transfer, among others. Due to increasing interest of using P2P systems for distributed computing, peer selection is taking relevance and several models have been proposed in the P2P literature. Yet, there are very few experimental studies for peer selection in P2P networks deployed in real large scale networks. In this work we present an experimental study that aims at revealing empirical information about the efficiency of selecting peer nodes in a P2P network when deployed in a real geographically distributed network. To this end, we have used a JXTA-based P2P platform deployed in Planet- Lab, a planetary scale infrastructure, and have empirically evaluated several peer selection models. Our experimental study showed that in order to achieve efficient P2P applications, appropriate selection model should be used according to the characteristics of the application. (http://doi.ieeecomputersociety.org/10.1109/ICPPW.2007.24 )
keyword : learning grids
Type de document :
Communication dans un congrès
International Conference on Parallel Processing (ICPP-2007), 2007, United States. pp.6, 2007, 〈10.1109/ICPPW.2007.24〉
Liste complète des métadonnées

https://telearn.archives-ouvertes.fr/hal-00190815
Contributeur : Jerome Zeiliger <>
Soumis le : vendredi 23 novembre 2007 - 08:58:18
Dernière modification le : vendredi 27 mars 2015 - 14:57:18

Lien texte intégral

Identifiants

Collections

Citation

Fatos Xhafa, Leonard Barolli, Raul Fernandez, Thanasis Daradoumis. An Experimental Study on Peer Selection in a P2P Network over PlanetLab. International Conference on Parallel Processing (ICPP-2007), 2007, United States. pp.6, 2007, 〈10.1109/ICPPW.2007.24〉. 〈hal-00190815〉

Partager

Métriques

Consultations de la notice

134