A control layer for a Peer-to-Peer middleware using behavior semantics

The increasing demand on networked virtual reality environments (VRE) is putting pressure to adopt scalable architecture capable to support new participants and deliver new services. Client/Server (C/S) architecture is showing different limitations especially in terms of scalability. Proposed solutions are based on a migration to a decentralized Peer-to-Peer (P2P) architecture that still lacks the control layer previously hosted by the server role. On the other hand, distributed data doesn't share same significance for maintaining synchronization between environments. Thus, controlling what data to distribute and when by defining interest on selected data according to certain criteria help in building generic strategies for the distribution process of VRE. In this paper, we propose a control layer for a P2P architecture based on behavior semantic-knowledge of VRE as a dynamic selection criteria for data of interest. Our proposal is presented using a semantic-rich framework named MASCARET distributed using a pure P2P middleware named Data Distribution Service (DDS).