Abstract — Peer selection for query routing is a core task in peer-to-peer networks. Unstructured peer-to-peer systems (like Gnutella) ignore this problem, leading to an abundance of network traffic. Structured peer-to-peer systems (like Chord) enforce a particular, global way of distributing data among the peers in order to solve this problem, but then encounter problems of network volatility and conflicts with the autonomy of the peer data management. In this paper, we propose a new mechanism, INGA, which is based on the observation that query routing in social networks is made possible by locally available knowledge about the expertise of neighbors and a semantics-based peer selection function. We validate INGA by simulation experiments with different data sets. We compare INGA with competing peer selection mechanisms on resulting parameters like recall, message gain or number of messages produced.

Although the ant metaphor has been successfully applied to routing of data packets both in wireless and fixed networks, little is known yet about its appropriateness for search in peer-to-peer environments. This paper presents SemAnt, a distributed content-based routing algorithm based on the Ant Colony Optimization meta-heuristic and adapted for deployment in peer-to-peer networks. Under the assumption that content is annotated according to a taxonomy, it is possible to determine the hierarchical relationships between queries, and to exploit this information to improve the routing process. Our results show that using taxonomies enhances search performance in peer-to-peer networks. The degree of enhancement is highly dependent on the content distribution in the network.

Abstract. Knowledge sharing in a virtual organization requires a knowledge life cycle including knowledge provisioning, terminology alignment, determination of resource location, query routing, and query answering. In this talk we focus on the issue of determining a relevant resource in a completely decentralized setting such as necessitated by peer-to-peer knowledge management in virtual organizations. Requirements for this task include, e.g., full autonomy of peers as well as full control over own resources and therefore preclude prominent resource location and query routing schemes such as distributed hash tables. In order to tackle given requirements we use a resource location and query routing approach that exploits social metaphors of topical experts and experts ’ experts as well as semantic similarity of queries and information sources. The approach has been fully tested in simulation runs and partially implemented in the system Bibster

This paper derives the optimal search time and the optimal search cost that can be achieved in unstructured peer-topeer networks when the demand pattern exhibits clustering (i.e. file popularities vary from region to region in the network). Previous work in this area had assumed a uniform distribution of file replicas throughout the network with an implicit or explicit assumption of uniform file popularity distribution whereas in reality, there is clear evidence of clustering in file popularity patterns. In this paper, we provide mechanisms for modeling clustering in file popularity distributions and the consequent nonuniform distribution of file replicas. We provide results for the search time in such networks for both random walk and flooding search mechanisms. The potential performance benefit that the clustering in demand patterns affords is captured by our results. Interestingly, the performance gains are shown to be independent of whether the search network topology reflects the clustering in file popularity. We also provide the relation between the queryprocessing load and the number of replicas of each file for the clustered demands case showing that flooding searches may have lower query-processing load than random walk searches in the clustered demands case.

We consider the network structure and query processing capabilities of social communities like bookmarks and photo sharing communities such as del.icio.us or flickr. A common feature of all these networks is that the content is generated by the users and that users create social links with other users. The evolving network naturally resembles a peer-to-peer system, where the peers correspond to users. We consider the problem of query routing in such a peer-to-peer setting where peers are collaborating to form a distributed search engine. We have identified three query routing paradigms: semantic routing based on query-to-content similarities, social routing based on friendship links within the community, and spiritual routing based on user-to-user similarities such as shared interests or similar behavior. We discuss how these techniques can be integrated into an existing peer-to-peer search engine and present a performance study on search-result quality using real-world data obtained from the social bookmark community del.icio.us. 1

Given a set of peers with overlapping interests where each peer wishes to keep track of new documents that are relevant to their interests, we propose a self-organizing peerto-peer document-tracking network based on common interest profiles. The goal of a document-tracking network is to disseminate new documents as they are published. Peers collaboratively share new documents of interest with other peers. There is no explicit profile exchange between peers and no global information available. We describe a strategy for peers to discover the existence of other peers and learn about their interests locally, based on information carried in the document metadata that propagates through the network. Peers are connected based on their observed common interests. We compare our proposed common interest strategy with a randomly connected network. The experimental results, based on simulated environment using the ACM digital library metadata, demonstrate that the proposed strategy gives the best dissemination performance. We also demonstrate that our self-organizing networks follow the characteristics of social networks.

In this paper we address the query routing problem in peerto-peer (P2P) information retrieval. Our system builds up on the idea of a Semantic Overlay Network (SON), in which each peer becomes neighbor of a small number of peers, chosen among those that are most similar to it. Peers in the network are represented by a statistical Language Model derived from their local data collections but, instead of using the non-metric Kullback-Leibler divergence to compute the similarity between them, we use a symmetrized and “metricized ” related measure, the square root of the Jensen-Shannon divergence, which let us map the problem to a metric search problem. The search strategy exploits the triangular inequality to efficiently prune the search space and relies on a priority queue to visit the most promising peers first. To keep communications costs low and to perform an efficient comparison between Language Models, we devise a compression technique that builds on Bloom-filters and histograms and we provide error bounds for the approximation and a cost analysis for the algorithms used to build and maintain the SON.