Expertise-Based Peer Selection

6.1 Introduction Peer-to-Peer systems are distributed systems without centralized control or hierarchical organization, in which each node runs software with equivalent functionality. A review of the features of recent Peer-to-Peer applications yields a long list: redundant storage, permanence, selection of nearby servers, anonymity, search, authentication, and hierarchical naming. Despite this rich set of features, scalability is a significant challenge: Peer-to-Peer networks that broadcast the queries to all peers do not scale-intelligent query routing and network topologies are required to be able to route queries to a relevant subset of peers. In this chapter we give an overview and an evaluation of the model of expertise based peer selection as proposed in [4] and how it is used in the Bibster system 18. In this model, peers use a shared ontology to advertise semantic descriptions of their expertise in the Peer-to-Peer network. The knowledge about the expertise of other peers forms a semantic overlay network, independent of the underlying network topology. If the peer receives a query, it can decide to forward it to peers about which it knows that their expertise is similar to the subject of the query. The advantage of this approach is that queries will not be forwarded to all or a random set of known peers, but only to the ones that have a good chance of answering it. The organization of the sections in this chapter is as follows: In the next section, we give a small overview of related work in the domain of Semantic Overlay Networks. In section 3 we provide our generic model on expertise-based peer selection. In section 4, we instantiate the generic model with the Bibster case-study. In section 5, we show simulation experiments and their results on the selection method. Section 6 shows the results of an evaluation study on the Bibster application which was installed on different machines of interested people. Section 7 shows a comparison between the simulation results and the results obtained from the field study. Section 8 concludes our work.