Efficient semantic search in peer-to-peer systems

While many P2P-based data sharing applications have been deployed, most of them just support semantics-free and coarse granularity (file level) sharing. Moreover, they are not efficient, either from user's view on service quality or in terms of resources utilization in the systems. In this paper, we present our solutions to support efficient semantic-based search in unstructured P2P systems. We propose a scheme to categorize and manage data in a peer based on the vector-space model. We also propose a scheme that allows peers with similar content to be clustered together. Finally, we examine an adaptive scheme to route queries. We conduct an extensive experimental study to evaluate the effectiveness of our solutions and obtain promising results.