TRIBLER: a social-based peer-to-peer system

It is observed that users tend to like and follow a particular type music/movies based on their social interests. Users like music of particular artists, album; movies/TV shows of a particular actor, director, genre, etc. Generally, peer-to-peer networks do not use this vital information to perform effective search using the metadata of media files. This information can be used to create social profile of users and cluster the peers based on their interests as reflected in the social profiles. This clustering of users over a peer-to-peer network will help in finding the users their interested files more easily. Also, peer-to-peer networks usually flood the entire network to search a particular file. But, due to clustering of peers based on their social interest, there is no need to flood the entire network to find a particular file, as most of the times the required file will be found in the cluster itself, which will in turn reduce the network overhead. We simulated this system using JXTA to connect the peers after clustering over P2P network. The peers with similar social profiles were connected via a common communication channel and therefore the results were better than the normal P2P network.

Bases de Données …, 2010

Dans cet article, nous proposons P2Prec, un service de recommandation pour systèmes de gestion de contenus P2P, qui exploite les données sociales des utilisateurs. L'idée est de recommander à l'utilisateur des documents de qualité en utilisant les notes données par les amis (ou amis d'amis) experts du sujet. Pour gérer les données sociales, nous nous appuyons sur des descriptions FOAF. P2Prec a une architecture P2P hybride, indépendante du système de gestion de contenus P2P, qui exploite l'efficacité d'une DHT pour indexer les fichiers FOAF et la robustesse du gossip pour disséminer les sujets d'expertise entre amis. Dans notre évaluation expérimentale, avec les données de CiteSeer, nous montrons que P2Prec permet d'obtenir un rappel maximum, avec de très bonnes performances. De plus, il améliore le rappel et la précision des réponses par un facteur 2 par rapport aux solutions centralisées.

2008 IEEE 17th Workshop on Enabling Technologies: Infrastructure for Collaborative Enterprises, 2008

In Peer-to-Peer (P2P) file sharing systems, peers have to choose the files of interest from a very large and rich collection of files. This task is difficult and time consuming. To alleviate the peers from the burden of manually looking for relevant files, recommender systems are used to make personalized recommendations to the peers according to their profile. In this paper, we propose a novel recommender scheme based on Peers' Similarity and Weighted Files' Popularity. Simulation results confirm the effectiveness of the Symmetric Peers' Similarity with Weighted File Popularity scheme in providing accurate recommendations, this way, increasing peers' satisfaction and contribution since peers will be motivated to download the recommended files and serve other peers meanwhile.

2012 IEEE Symposium on Computers and Communications (ISCC), 2012

Content management appears as an essential requirement in order to deploy enhanced P2P networks. In P2P networks, the content that is requested by the query node could be located at different locations/nodes; therefore an efficient search mechanism is required. The proposed search algorithm in this paper, called socP2P, relies on peers' social relationships (friendships, shared interests, shared background and experiences) to improve the content discovery compared to similar solutions. With socP2P nodes can improve searches by using knowledge gained by 'overheard information' during their stay in the network. In addition, our algorithm exploits peers' common interests, friendships, and capability of memorizing experiences (received and routed queries) by them. Simulation results show that socP2P is able to achieve a high success rate, low delay and low overhead. We have verified that our algorithm is not only useful in finding popular contents in the network but also good enough to locate rare files. The obtained results reflect that exploiting social information in P2P networks leads to a more efficient content search mechanism.

2012

P2PShare est un système Pair à Pair (P2P) pour le partage à grande échelle de données probabilistes qui s'appuie sur un système de recommandation basée sur le contenu et l'expertise de chaque utilisateur. Il est conçu pour gérer des données probabilistes et déterministes dans un environnement P2P. Il fournit un environnement flexible pour l'intégration de sources hétérogènes, et prend en compte les aspects sociaux pour découvrir des résultats de haute qualité pour les requêtes en privilégiant les données des amis (ou des amis d'amis), qui sont les experts sur les sujets liés à la requête. Nous avons mis en place un prototype de P2PShare à l'aide du logiciel Shared-Data Overlay Network (SON), une plate-forme de développement open source pour des réseaux P2P utilisant les services web, JXTA et OSGi. Dans cet article, nous décrivons les principaux services de P2PShare, par exemple, la diffusion des sujets d'intérêt entre les amis, les requêtes sur les contenus et les requêtes probabilistes sur les jeux de données.

2003

The objective of this project is to simulate a Peer-to-Peer type of environment with the JADE multi-agent system platform to investigate the use of social networks to optimize the speed of search and to improve quality of service in the Peer-to-Peer environment. Our project uses the Gnutella protocol as a starting point. The Gnutella protocol broadcasts messages for searching files. This message passing generates much traffic in the network. This degrades the quality of service. We propose a model where each peer has a "friends list", for each category of interest. Once peers generate their "friends list", they use these lists for searching files in the network. The model has been implemented and some initial experiments have been performed.

2010 10th IEEE International Conference on Computer and Information Technology, 2010

In this paper, we propose a novel recommender framework for P2P file sharing systems. The proposed recommender system is based on user-based collaborative filtering technique. We take advantage from the partial search process used in partially decentralized systems to explore the relationships between peers. The proposed recommender system does not require any additional effort from the users since implicit rating is used. To measure the similarity between peers, we investigate similarity metrics that were proposed in other fields and adapt them to file sharing P2P systems. We analyze the impact of each similarity metric on the accuracy of the recommendations. Files' recommendations will increase users' satisfaction since they will receive recommendations on files that they prefer.

Proceeding of the eleventh ACM SIGKDD international conference on Knowledge discovery in data mining - KDD '05, 2005

We use knowledge discovery techniques to guide the creation of efficient overlay networks for peer-to-peer file sharing. An overlay network specifies the logical connections among peers in a network and is distinct from the physical connections of the network. It determines the order in which peers will be queried when a user is searching for a specific file. To better understand the role of the network overlay structure in the performance of peer-to-peer file sharing protocols, we compare several methods for creating overlay networks. We analyze the networks using data from a campus network for peer-to-peer file sharing that recorded anonymized data on 6,528 users sharing 291,925 music files over an 81-day period. We propose a novel protocol for overlay creation based on a model of user preference identified by latent-variable clustering with hierarchical Dirichlet processes (HDPs). Our simulations and empirical studies show that the clusters of songs created by HDPs effectively model user behavior and can be used to create desirable network overlays that outperform alternative approaches.

Future Generation Computer Systems, 2014

• Conduct extensive studies on social information from 384,494 Facebook users.

Expert Systems With Applications, 2008

Recommender system is a popular technique for reducing information overload and finding digital contents that is most valuable to users. However, most recommender systems are based on a centralized client-server architecture in which servers and clients represents contents providers and users respectively. The existing recommender systems depend on contents providers and give a number of disadvantages to users. Therefore, we propose a recommender system based on a distributed P2P architecture that has originated with user-oriented principle rather than business itself. The proposed system consists of fully functioning personal recommender agents that automatically select neighbors and recommend contents. The agents learn user preference from users' content usage without requiring users' explicit ratings. We believe that the suggested P2P based recommender system should provide the users with more qualified recommendations, while it reduces the effort and time of users.