Community Oriented Neighboring Feedback for Efficient End-to-End Delay-Sensitive MP2P Streaming via Temporal Relay-Based Replication

Abstract

Next generation peer-information systems are envisioned to be characterized by an invisible and ubiquitous lookup application support which should be easily accessible by users in a transparent and location-independent reliable fashion. Therefore, the traditional end-to-end delays should be bounded and minimized in order to offer significant increase in performance and reliability in today's Mobile Peer-to-Peer (MP2P) systems. Many parameters that affect the end-to-end delay for efficient collaborative streaming, deal with the temporal characteristics (connectivity and mobility constraints, etc). These parameters are significantly degrading the end-to-end communicational survivability and resources' availability. In this work a scheme for collaborative replication is proposed using community oriented relay regions for MP2P streaming. The proposed model considers a combined community oriented neighboring feedback approach with selective delay-centric replication taking into account the social interactions for assigning available resources to nodes that are communicating via intraclustered links. The collaborative streaming is achieved through the outsourcing file chunk policy for N-hop replicated sources and the combined community oriented model. The presented scheme is thoroughly evaluated through experimental simulation which takes measures for the end-to-end reliability and the latency of the broadcast schedule when delivering delay-sensitive MP2P streams.