铁通业务监控平台数据采集系统的设计与实现

【作者】 王芳；

【导师】 叶德建；

【作者基本信息】 复旦大学 ， 软件工程， 2011， 硕士

【摘要】 中国铁通北京分公司由于设备多、品牌杂、人手少的原因,存在设备监控效率低下的问题,而监控平台具有体积小,稳定性高,可扩展性强,可适用于多平台等优点,具有很高的IT运维管理价值,非常适合提高铁通设备的有效监控能力。本文主要是通过Agent-server模式来实现数据采集功能,从而实现设备监控。Agent是轻量级代理服务程序,被部署在单独代理服务器上,用于接收Server的数据采集需求,并为Server提供采集数据。Server是该业务监控平台的核心管理模块,被部署在被监控的服务器上,通过收集Agent采集的被监控对象的运行时数据,并将这些数据转换成统一的模型数据,进行必要的计算、汇总,以及规则计算后存储,从而实现对IT资源的事件管理、性能管理。本文重点围绕Windows下的采集系统进行设计和说明。首先对于整个监控平台及数据采集系统的构架进行了设计,在此基础上对于数据采集系统的各个组成模块进行了详细的说明,并为此设计了多个数据采集程序,同时采用Python这个目前使用非常普遍的网络开发语言进行了实现,提供了对应的源代码。最后,根据铁通公司的实际业务情况,对于数据采集系统进行了网络配置,在程序的单元测试时,采用白盒法进行测试。同时根据测试结果,从功能、安全性、数据一致性等方面评估了该数据采集系统,发现测试结果基本达到了期望,对于提高铁通的监控效率具有较大的帮助。更多还原

【Abstract】 Because of lots of equipments, miscellaneous brand and less manpower in China Tietong Beijing Branch, there is the problem of low efficiency in equipment monitoring, then monitoring and control platform has the characters, such as small size, high stability, scalability and suitable for multi-platform, etc, so it has high value in IT operation and is very adapted for improving efficient about monitor and control.This article mainly realizes data acquisition through the Agent-server mode for monitoring equipments. Agent is a lightweight proxy service program and is deployed on separate proxy server to receive data collection needs of the server and provide collected data. Server is the core of the business management module of this monitoring and control platform, it is deployed on the server being monitoring, Through collecting real-time data of object being monitoring that is gathered by Agent, the Server transforms these data into unified model data and perform necessary calculations, summary and store, then achieve event management and performance management for IT resource.This article focuses on the collection system design in windows. Firstly, the whole frame of monitoring and control platform and data collection system are design, on the basis of the frame various components of data collection system is described in detail, furthermore, several data collection procedures are designed and realized by Python that is very popular network development language. Finally, according to the actual business situation, network configuration of data collection system is realized for testing the system, White-box method is utilized when unit test for Procedure, then according to test results of functionality, security and data consistency, the data collection system are evaluated. From the evaluated result, the system basic meets expectations and has a large help for improving the efficiency.更多还原