Design And Implementation Of Network Traffic Control System For Three-level Network Based On OPC And Service Classification

With the rapid development of computer network technology, network haspenetrated into all aspects of social life, and various industries erect different scale LANfor office. Part of them with a typical three hierarchical topology is named Three-levelNetwork in short. How to optimize the network configuration, improve networkperformance and guarantee the quality of service of the network application through thecharacteristics of three-level network, have become an urgent problem to be solved.Based on the study of OPC technology and present network applicationclassification techniques, combined with the characteristics of applications andterminals in three-level network, for overcoming the defects of exiting networkmanagement models, a network traffic control system for three-level network based onOPC and service classification is designed and implemented in this thesis.First of all, on the basis of analyzing the characteristics o f three-level network andOPC technology, the traffic control system based on OPC which is used ascommunications protocol for network terminals, applies packet blocking technology forflow control to reduce the pressure of the network equipment. The netwo rk applicationsare divided into key and non-key application, and the web application performancemetrics defined by ITU-T are specified for guaranteeing the key applications’ quality ofservice and controlling flow.Next, after a general study on network application classification and multipleclassifiers integration technologies, a dynamic self-adapting and self-feedbackintegration framework with multiple classifiers is designed for network applicationclassification. The framework sets a distributor and a feedback-unit for dynamiccombination of multiple classifiers. The distributor is used for selecting base classifier,and the feedback-unit is used for improving the distribution. The experiment resultsshow that, the framework can reduce the number of base classifier used and theconsumption of the performance with high accuracy of classification.Finally, a three-level network traffic control system is designed and developed withthe dynamic self-adapting and self-feedback integration framework with multipleclassifiers. By using hierarchical distributed architecture based on OPC technology anda hierarchical interactive architecture, the system has implemented to collect networkflow and identify network application, detect network status and key application’srequirements of quality of service, analysis network performance and choose trafficcontrol strategies, control network flow of terminals. The experiment results show thatthe system can better analysis network performance, accurately identify the key networkapplication, effectively detect key application’s requirements of quality of service andguarantee the quality of service of key application by controlling flow.