Network packet loss rate boundary characterization method under multiple-data stream fusion mechanism

Abstract

The invention relates to a network packet loss rate boundary characterization method under a multiple-data stream fusion mechanism. The method is characterized by comprising the steps of designing an arrival model, a service model and a corresponding boundary function on the basis of the distribution characteristics of flow backlog, modeling a coding node based on a fusion service mechanism, deriving the boundary function and the arrival curve of the node in a phased manner, and establishing the corresponding relationship between the system packet loss rate boundary performance on one hand and the system cache capacity, the data flow arriving the node, the service ability and the quality of service demand on the other hand based on the algebra theory under both conditions of full-fusion opportunity and no-fusion opportunity, thus getting the packet loss rate boundary performance under multiple-data stream fusion service. According to the invention, the arrival process and the service process are modeled based on a cumulative flow center arrival curve and a strict random service curve for the first time, and the behaviors and characteristics to which the two different random processes pay close attention are embodied. The method is more practical, and meets the needs of operating units for network configuration.

Description

technical field [0001] The invention relates to a method for describing the boundary of network packet loss rate, in particular to a method for calculating the boundary of packet loss rate based on multiple data streams in a fusion environment, and belongs to the field of electronic communication. Background technique [0002] As we all know, in the field of information communication, network coding based on linear XOR allows one or more nodes in the network to fuse multiple packets received by different paths into a single coded data packet through linear information processing. Continuous transmission, this idea has completely subverted the previous data transmission technology, thereby completely changing the traditional information processing concept based on the store-and-forward mode, and can achieve the theoretical maximum throughput calculated in network information theory, so it has become a popular technology in industry and academia. It clarifies the research dire...

AI Technical Summary

Problems solved by technology

At present, there are more or less defects in the solution to the problem of packet loss rate: (1) It is more intuitive to use the Markov model to analyze, but it is difficult to calculate the exact value because of the explosion of the state space, and the fused data flow The basic properties of Markov are no longer maintained (for example, no memory, multiple input streams are Poisson processes), and the related packet loss rate problems are calculated based on such properties; (2) use the minimum plus algebra To describe the packet loss rate boundary, the general solution result can be obtained, but the boundary is too loose, and it is difficult to describe the performance after multiple hops. The conclusion given is a resource allocation scheme in the most conservative case (worst case) , even though the boundary of the packet loss rate in this case is very tight, but because the probabi

Images