Research On Large-scale P2P Network Streaming Media Interaction System

With the development of network bandwidth and streaming media technology,streaming media interaction is experiencing explosive growth,and real-time streaming media interaction will become the mainstream application scenario of the future Internet.The large-scale streaming media interaction system brings unprecedented difficulties to the network software and hardware architecture,bandwidth,operation and maintenance.In the face of large-scale streaming media interaction scenarios,the traditional client/server architecture is difficult to cope with,and the Peer-to-Peer architecture is inevitable.As a real-time network application,the interactive streaming system requires high startup delay,end-to-end delay,and playback continuity during transmission,directly affecting the user experience.This paper focuses on the construction theory and technology of large-scale P2 P network streaming media interaction system,and studies two key factors that affect the real-time and continuity of large-scale P2 P network streaming media interaction system,namely topology structure and data scheduling.The main work of this paper is as follows:Firstly,the P2 P network characteristics,model structure and related applications are analyzed,including the organization and routing between nodes,the query between nodes and the joining and exiting of nodes;the research on the file format and related transmission protocol of streaming media technology,and The related system model of P2 P streaming media technology is analyzed to study the characteristics and application scope of P2 P network technology,and the advantages and disadvantages of existing system models.Secondly,the network topology is studied on the large-scale P2 P network streaming media interaction system.For the detailed analysis of the organization and routing between nodes,the query strategy between nodes and the joining and exiting mechanism of nodes,a hybrid hierarchical tree network combining method based on Kademlia protocol is proposed.The physical distance is introduced based on the Kademlia protocol.The combination of logical distances ensures that network routing is more efficient.Thirdly,research and design the data scheduling strategy between nodes.A node capability evaluation method based on node uplink bandwidth,online duration,contribution,and the like,and a corresponding super node selection mechanism are established,so that the super node is closer to the publishing source node,and the distribution efficiency of the streaming media data block is improved.In addition,in order to ensure the continuity of streaming media data during transmission and the load balancing between nodes,the urgency of the data block transmission process is based on real-time monitoring and measurement of the state of the data block between the video playback point and the push pointer.The collaborative selection of the push and pull operations of the streaming media data block in the corresponding data scheduling process is guided.The experiment proves the effectiveness of the method.Finally,the system simulation and performance test of the proposed scheme are carried out on the simulation platform Over Sim,including three measurement indicators: startup delay,end-to-end delay and video distortion rate.Compared with the existing schemes,the scheme designed in this paper reduces the startup delay,end-to-end delay and video distortion rate by at least 11%,16% and 12% respectively compared with the existing schemes,indicating the scheme studied in this paper.And technology has achieved the desired goals and can be applied to large-scale streaming media interaction systems.