蒋方强

摘 要：随着企业各类业务的持续增长，IT设施也在不断完善和扩建。网络线路、网络设备和服务器等基础设备不断增加，整个网络结构变得异常庞大且复杂，一定程度上增加了监控和管理的难度。许多关键的业务系统往往包括网络线路、防火墙、前端应用服务器、负载均衡设备、中间件、数据库和存储等各类软硬件设施，在传统管理手段下，对于多段网络、多层应用的架构，运维管理关联度不够高，导致在问题发生时，分析定位相对比较困难，排障速度相对较慢。部署面向应用的网络流量分析系统，可以建立统一的网络流量监测分析的底层平台，自动梳理应用系统网络访问路径，同时建立起面向应用的网络服务视图，实现全网流量监测，并且根据流量特征实现快速故障预警与故障诊断，更加科学有效地的管理IT设施，确保业务系统稳定运行。

关键词：面向应用;流量特征;网络流量分析;流量监测

中图分类号：TP393.18文献标识码：A文章编号：1003-5168（2021）10-0009-04

Application Oriented Traffic Visualization System of the Whole Network

JIANG Fangqiang

（Shanghai Pudong Development Bank，Shanghai 200233）

Abstract： With the continuous growth of all kinds of business， it facilities are constantly improving and expanding. With the increase of network lines， network equipment and servers， the whole network structure becomes very large and complex， which increases the difficulty of monitoring and management to a certain extent. Many key business systems often include network lines， firewalls， front-end application servers， load balancing devices， middleware， databases， storage and other software and hardware facilities. Under the traditional management methods， for the multi segment network and multi-layer application architecture， the operation and maintenance management relevance is not high enough， which makes it relatively difficult to analyze and locate problems， the speed of obstacle removal is relatively slow. Deploying the application-oriented network traffic analysis system can establish a unified underlying platform for network traffic detection and analysis， automatically comb the network access path of the application system， establish an application-oriented network service view， realize the visualization of the whole network traffic， and realize rapid fault warning and fault diagnosis according to the traffic characteristics， more scientific and effective management of it facilities to ensure the stable operation of business systems.

Keywords： application oriented;flow characteristics;network traffic analysis;flow monitoring

隨着企业各类业务的持续增长，IT基础设施设备数量不断增加，网络、系统、应用的关联性日益密切，整个结构变得异常庞大且复杂。在传统管理手段下，对于多段网络、多层应用的架构，运维管理关联度不够高，导致在发生问题时分析定位相对比较困难，排障速度较慢。在运维工作中，应用、系统和网络管理人员往往需要付出额外的工作量，依靠技术积累和经验进行人工分析、定位、关联以应对多层面联动的异常及故障[1]。……