Non-state end-to-end constraint entrance permit control method for kernel network

Abstract

The present invention is state-less end-to-end multi-constraint admittance control method of kernel network and belongs to the field of Internet technology. The method features that the entire Internet is divided into kernel network and edge network in the basis of no-connection gradually jumping QoS route, and the kernel network maintains its own QoS information and uses QoS route without maintaining each flow state; and that the admittance control is completed in the source gateway of the edge network via the matching of all nodes in the end-to-end path, with the source gateway completing the flow counting and detecting function. The method utilizes the network resource reasonably and raises the utilization of the network links.

Description

technical field [0001] The stateless end-to-end multi-constraint admission control technology of the core network belongs to the field of Internet service quality control, and particularly relates to the admission control technology with multiple service quality parameters. Background technique [0002] How to provide different quality-of-service (QoS) guarantees for applications is an important problem faced by the Internet. In order to allocate and utilize network resources reasonably, admission control has become one of the essential core mechanisms. Admission control needs to have good scalability while providing a reasonable network resource allocation method. The existing admission control research schemes can be divided into three categories: (A) based on bandwidth proxy, (B) based on integrated services, (C) based on actual measurement, and their research results have the following shortcomings: (1) violating the Internet connectionless, Hop-by-hop routing nature w...

AI Technical Summary

Problems solved by technology

Existing admission control research schemes can be divided into three categories: (A) based on bandwidth proxy, (B) using integrated services, (C) based on actual measurement, and their research results have the following shortcomings: (1) violate the Internet connectionless, The essence of hop-by-hop routing without resource reservation (such as class AB); (2) the number of convection flows and the scalability of the network scale are poor (such as class AB); (3) the admission control delay is large (such as class C); ( 4) It is impossible to realize the control of different service quality parameters (such as BC class)

Therefore, there is still a certain distance between the existing research results and the actual Internet

Images