Network System Design And Implementation For Experiments Management And Analysis

With the development of military information construction and the complication of confrontation environments in battlefields,the communication network management of private network has become the basic work for building professional information teams.How to effectively coordinate information exchange among various systems in the test project,how to efficiently guarantee the smoothness of the communication network links in the private network,and how to demonstrate the status information of the network link nodes in experiments have become the main research content for current military experiments management.Based on the research and analysis of the given experimental workflow,this thesis aims at solving such problems as the lack of automatic flow for message data interaction,the poor visualization of resource topology information,and the lack of display and analysis function for fault warning.This thesis summarizes and analyzes the original system,designs and implements a new network system for experiments management and analysis based on B/S architecture.The main work and contribution are summarized as the following.1.Module design and implementation for system message interaction.This thesis analyzes the mode of data and message state interaction between the system,the upper management control system and the lower layer simulation subnet system,uses RTI Connext DDS data distribution middleware for module design and implement about system information interaction via the work mode for request response and publish / subscribe.2.Module design and implementation for key information visualization.In this thesis,the second-development of customized interface based on ExtJs front-end frame is carried out,where the supervised fault information in the process of experiments is summarized and displayed based on ECharts technology,and the communication network two-dimensional topology planning function and three-dimensional comprehensive situation display function are realized by means of TWaver graphics development tool.3.SVM-based model design and implementation for system fault alert.In this part,machine learning in introduced for system fault alert model construction,and a SVM-based classifier is trained for system fault diagnosis.Here,a course-to-fine grid searching strategy is applied for super-parameter optimization.Experiments show the effectiveness of the proposed model.In conclusion,the system standardizes the flow of message interaction in experiment,finishes the friendly interface development for resource management,fault alarm management,topology display.The final system can monitor the resource status information and link state information in the process of test in real time,and supply fault alarm information by means of SVM classifier.