An implementation method for large-scale stream media transmission structure

Abstract

A method realizes a large-scale stream media transmission structure and has five steps. Firstly, a started retransmission node controller sends a heartbeat message to determine the type of each media retransmission node; secondly, each media retransmission node sends a heartbeat message to determine the distribution status in each multicast island; thirdly, each multicast island elects an aggregation node, which synthesizes the topological structure of multicast islands and the information of nodes that are then sent to the retransmission node controller, which, therefore, knows the global topological structure; fourthly, as for every newly admitted user terminal, the retransmission node controller selects a media retransmission node for the user terminal by the method of automatic accessof users; finally, after the access of the user terminal, for each newly established session, each neighboring multicast island elects a communication node or an agent node, which takes charge of thedata intercommunication between each neighboring multicast island and the main multicast island, and the data message sent by the user terminal is correctly transmitted to other user terminals in thesession according to the method of multimedia data retransmission between each communicated node and each agent node. The present invention effectively increases the scale and stability of system access and the interactive quality among different network operators.

Description

technical field [0001] The invention relates to the application of a large-scale multimedia network, in particular to a method for realizing a large-scale real-time streaming media transmission structure. Background technique [0002] The amount of data transmission in the video conferencing system is large. Under the traditional central server / client model, due to the limited bandwidth and processing capacity of the server, the system scale and scalability are limited. IP multicast can save network bandwidth to the greatest extent, but due to technical and commercial reasons, multicast has not been widely deployed, and a large number of IP multicast islands are distributed in the Internet. Therefore, breaking through the traditional server / client transmission model and making full use of the transmission capability of IP multicast islands will effectively save network bandwidth and server processing capabilities. Application-layer multicast utilizes terminals or forwarding...

AI Technical Summary

Problems solved by technology

Application-layer multicast utilizes terminals or forwarding nodes deployed in the network to implement multicast services at the application layer and can easily utilize the multicast capabilities of IP multicast islands. In traditional application-layer multicast, Scribe, Bayeux, and CAN are based on distribution Multicast is realized by hash table, its logical neighbor relationship cannot reflect the physical proximity, the transmission delay is relatively large, and it is not suitable for real-time streaming media transmission; Narada, HMTP, Yoid, ALMI, NICE, HostCast, they all rely on the end system to achieve Multicast, transmission delay and system stability are difficult to guarantee

Constructing an overlay network by strategically deploying forwarding nodes in the network is the development direction of large-scale and high-quality group communication. In the application layer multicast application based on the overlay network, RMX is suitable for reliable application layer multicast, Overcast, OMNI is suitable for single data source applications, and Scattercast emphasizes the wide range of applications, which cannot meet the needs of real-time streaming media transmission; and these overlay network structures realize the correct distribution of data packets by building and maintaining multicast distribution trees, but the group The construction and maintenance of the broadcast distribution tree consumes a large amount of control overhead and node memory, which limits the system scale and affects system stability

Images