An end-to-end voice communication node addressing method based on two-layer structure

Abstract

The invention discloses an end-to-end voice communication node addressing method based on a double-layer structure. When each node participating in the communication is connected to the network, it is logically divided into two layers, and the upper layer is composed of a super node through a distributed hash table mechanism. To connect, the lower ordinary node selects a super node of the upper layer as the host node, and this super node implements the distributed hash table mechanism publishing and searching of ordinary node information. When a node accesses, it will distinguish super nodes, backup super nodes and ordinary nodes according to their own network address properties, network bandwidth capabilities, computing processing capabilities, and storage capabilities. The distributed hash table mechanism adopted applies the improved K-bucket structure, adding the IP address and port number information of the connected super node. In the process of node joining and addressing, super nodes and ordinary nodes will be implemented according to different steps. The invention can effectively improve the addressing efficiency of the voice communication node.

Description

technical field [0001] The invention relates to the field of voice communication, in particular to an end-to-end voice communication node addressing method based on a double-layer structure. Background technique [0002] Traditional network voice communication requires the help of a central addressing server to address and locate the destination terminal. For example, users in the SIP voice communication protocol need to register their own information on the SIP server first. When the user makes a call, he needs to query the address information of the target user through the SIP server, and then through the corresponding signaling interaction, the streaming media communication channel can be established. . It can be seen from this that in the client / server mode, the server is responsible for the registration, query and maintenance functions of all end users, and the performance of the server directly affects the reliability and stability of the entire system. The server ha...

AI Technical Summary

Problems solved by technology

[0005] However, after an in-depth study of the operating mechanism of the Kademlia protocol, it can be found that the existence of the mismatch between the physical network and the logical network also has its rational factors. Any attempt to improve this mismatch will have problems.

The main reason is that the node distribution and search requirements in the Kademlia protocol have uniform characteristics. If the above-mentioned mechanism is used to improve the mismatch problem existing in the Kademlia protocol, the uneven distribution of nodes on the physical network will inevitably lead to uneven distribution of logical network nodes. characteristics, which violates the essence of end-to-end network design, making some popular nodes and even nodes with poor performance bear more load tasks

[0006] However, if the Kademlia protocol is improved so that logically similar nodes are also physically similar, the communication delay of the first few K buckets of nodes will be small, while the delay of other K buckets will be large, resulting in extremely low resource search efficiency. Inhomogeneity

Images