Research On Network Routing And Resource Allocation Based On Network Traffic Coupling Analysis

In recent years,users' network demands have increased and the Internet has grown rapidly.At the same time,the types of services and the volume of services carried by the network are also increasing.Different types of application traffic share limited network resources.Due to the uneven distribution and suddenness of user demand distribution,network congestion often occurs on some network elements.When the network is congested,there is a significant correlation between the qualities of the network traffic,called coupling in this paper.The complexity of the network structure and the diversity of network services make the form of this coupling behavior more compliccated.Further,the interaction between massive traffic in the network makes this coupling phenomenon fluctuate throughout the network and affects the network state.Studying the coupling behavior between service qualities of network traffic can better grasp the network state,which is of great significance for ensuring user service quality and optimizing network resource configuration.Researchers have done a lot of meaningful research and exploration in the field of network traffic,such as network traffic modeling,network traffic forecasting and so on.These studies mainly focus on the mining of the characteristics of network traffic,have not yet studied the coupling behavior between network traffic.Some of the research involves the correlation between network traffic,but only from the statistical or macroscopic study of this association behavior,there is no micro-action mechanism of coupling behavior.In view of the deficiencies of the existing research,this paper focuses on the coupling behavior between network traffic in complex network environment,mining the impact of coupling behavior on the network in depth,and providing guarantee for the reliable operation of the network according to the obtained results.Specifically,the main work of this paper is as follows:(1)Based on the theory of elasticity,we established the coupling behavior mechanism of the QoS of network traffic microscopically,which called the traffic-coupling model.The limitation of network resources determines that there must be a constraint relationship between network users.The specific performance is that there is a mutual influence between the QoS of network traffic.The traffic-coupling model abstracts the bandwidth coupling behavior by the interaction between the elastomers.The model solution results show that this coupling mode is equivalent to the bandwidth proportional differentiated service.In addition,the model establishes the corresponding relationship between communication system and physical elastic system,and provides a new perspective to understand network phenomena.Furthermore,based on the bandwidth solution of the model,this paper analyzes the efficiency and fairness of bandwidth allocation in proportional differentiated services by simulation.This paper explores how to set quality of service differentiation parameters for various network traffic,and provides an analysis method to determine proportional differentiation parameters.(2)The traffic-coupling model establishes the corresponding relationship between communication system and physical elastic system.This paper proposed "network strain energy" to measure the degree of network congestion as compared with elastic system.It is also the quantification of traffic coupling strength in the energy perspective.Considering the coupling behavior between network traffic,this paper proposes a network congestion oriented routing strategy,which aims at minimizing the strain energy increment of the network.The simulation results show that this routing strategy can not only reduce the coupling between network traffic,but also ensure the stability of the network state.At the same time,it has a good balance load capacity and improves the utilization of network resources.