A Routing Scheme for Content Localization in Peer-to-Peer Networks

Search and localization of video is a challenging problem in Peer-to-Peer (P2P) video streaming systems. Traditional approaches have been mainly based on a central server or flooding techniques. We present a decentralized scheme that exploits semantic for unstructured P2P systems. Our semantic approach classifies videos and peers into semantic groups. The proposed scheme uses semantic neighbors to route a semantic query. Semantic paths are discovered for every successful query and saved in a cache in the requesting peers. A Semantic Overlay Network (SON) is built from several semantic paths. Our semantic protocol reduces the search time and decreases the network traffic by reducing the number of messages in the system as compared to Gnutella. Using a SON, our protocol performs comparable to centralized systems as searching is restricted to only a specific peer or to a small number of peers.