Peer Selection in Peer-to-Peer Networks with Semantic Topologies

Peter Haase
Ronny Siebes
Frank van Harmelen

Peer-to-Peer systems have proven to be an effective way of sharing data. Modern protocols are able to efficiently route a message to a given peer. However, determining the destination peer in the first place is not always trivial. We propose a model in which peers advertise their expertise in the Peer-to-Peer network. The knowledge about the expertise of other peers forms a semantic topology. Based on the semantic similarity between the subject of a query and the expertise of other peers, a peer can select appropriate peers to forward queries to, instead of broadcasting the query or sending it to a random set of peers. To calculate our semantic similarity measure we make the simplifying assumption that the peers share the same ontology. We evaluate the model in a bibliographic scenario, where peers share bibliographic descriptions of publications among each other. In simulation experiments we show how expertise based peer selection improves the performance of a Peer-to-Peer system with respect to precision, recall and the number of messages.

(PDF paper, 325Kb)

@InProceedings{ICSNW04,
  author =       "P. Haase and R. Siebes and F. van Harmelen",
  title =        "Peer Selection in Peer-to-Peer Networks with Semantic Topologies",
  booktitle = 	 "Proceedings of the International Conference on Semantics in 
                  a Networked World ({ICNSW'04})",
  editor =       "Mokrane Bouzeghoub",
  year = 	 2004,
  series =	 "LNCS",
  publisher =	 "Springer Verlag",
  volume =       3226,
  pages =        "108-125",
  address = 	 "Paris",
  month = 	 "June",
  keywords = {Semantic Web},
  urlPaper = "http://www.cs.vu.nl/~frankh/postscript/ICSNW04.pdf"
}

<- Back to list of papers