Research, And Other Network-based Streaming Media Technology

With the rapid development of the Internet and multimedia technologies, streaming media applications over the Internet has been becoming more and more popular. However, due to the best-effort property of the Internet and the heterogeneity of networks and terminals, traditional Client-Server type of media streaming is never suitable for large scale media service. The rapid development of P2P technology makes it as one of the most disruptive tools for the construction of large-scale distributed system over Internet. Therefore, as one of the most perspective applications, P2P streaming system has become one of the hotspot of P2P research.This paper conducts in-depth research on several basic issues of P2P streaming system, including the design of P2P overlay network, optimizing the topology based on underlay network, managing member nodes, and scheduling dynamic data set between data sending nodes to obtain minimum delay, etc. Within the above issues, a project called AVStreamer is done to demonstrate a well organized P2P media streaming system.The main issues and contributions of this paper is as following:(1) Illistrated a new P2P overlay protocol SOLNP as the infrastructure of streaming application, which harness the heterogeneity of the nodes to maintain topology, and combines the issue of bandwidth efficiency, delay and reliability. Within such infrastructure a capacity-aware algorithm for topology maintenance is implanted, which can make full use of the overlay network to provide high efficiency, load-balance, low network radiance topology. Another key design involved in the protocol is a self-adapted topology optimizing algorithm, which makes the nodes select better quality partner nodes from the list to connect, and provides a stronger overlay network topology for the upper streaming media application.(2) Introduced an algorithm called SDSA to dynamic scheduling data distribution among sets of partner nodes, which relies on the resources allocation and computes the "data retrieving set" dynamically. With the consideration of node load and network delay, it controls the sequence of data sending nodes, which resulted in a balanced load network and low bandwidth cost.