WhatsUp: news from, for, through everyone - INRIA - Institut National de Recherche en Informatique et en Automatique Accéder directement au contenu
Communication Dans Un Congrès Année : 2010

WhatsUp: news from, for, through everyone

Résumé

WhatsUp (WUP) is a new form of electronic news. It is personalized and decentralized. Users receive news and have the ability to express their interest in it. This opinion, in turn, is used as an implicit and dynamic subscription scheme to filter and personalize future information. The system is peer-to-peer: no big brother company controls the news, and no central server makes it vulnerable to failures, censorship or attacks. At the heart of WUP lies the idea of collaborative filtering applied to the dissemination of news: people who liked the same news in the past might as well like the same news in the future: irrelevant news disappear by themselves. The idea is put to work through Beep: a biased epidemic dissemination (gossip) protocol that delivers news to interested users in a timely manner, despite jamming and churn. BEEP is dynamically parameterized on a per-user, per-news, and per-dissemination-hop basis. When compared to a classical epidemic dissemination protocol, BEEP has two key characteristics: orientation and amplification. Every user forwards the news of interest to a randomly selected set of users largely constituted by those who have similar interests (orientation). Moreover, the size of this set of users depends on the level of interest in the news itself (amplification).
Fichier principal
Vignette du fichier
demoWhatsUp-p2p10.pdf (55.9 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

inria-00515420 , version 1 (06-09-2010)

Identifiants

  • HAL Id : inria-00515420 , version 1

Citer

Antoine Boutet, Davide Frey, Rachid Guerraoui, Anne-Marie Kermarrec. WhatsUp: news from, for, through everyone. 10th IEEE International Conference on Peer-to-Peer Computing (IEEE P2P'10), Aug 2010, Delft, Netherlands. ⟨inria-00515420⟩
354 Consultations
260 Téléchargements

Partager

Gmail Facebook X LinkedIn More