Scalable CGM distribution system based on a semantic peer-to-peer network

This paper proposes a scalable CGM (Consumer Generated Media) distribution system that delivers newly published content in real time to the appropriate users whose preferences coincide with the content topic. In the proposed system, a semantic P2P network is constructed based on CAN (Content Addressable Network) to distribute content according to the similarity between user preferences and the content. The system realizes efficient and robust content delivery on a P2P network. It reduces network traffic traversed across ASs (Autonomous Systems) by reflecting the physical network location of peers in the topology of the P2P network. The system also handles frequent peer join and departure by preparing a few redundant paths in addition to the main ones. Performance evaluations are also discussed in this paper.