Semantic Overlay Networks for Social Recommendation in P2P

In P2P systems, nodes typically connect to a small set of random peers to query them, and they propagate those queries along their own connections. To improve that mechanism, semantic overlay networks influence those connections depending on the content of the peers, clustering peers in overlapped groups (Semantic Overlay Networks). We propose using ontologies for describing semantic information of both, the shared items and the user profile. Once the peers are grouped by semantic information, we can take advantage of that distribution to add some new functionalities as recommendation, using ontologies comparison and the same principle to create SONs: the users with similar files to us, will probably have files that we are interested in.