【作者】 周星；

【导师】 汪国安；

【作者基本信息】 河南大学 ， 应用数学， 2005， 硕士

【摘要】 随着计算机和通信技术的发展,计算机网络技术也得到了迅猛发展。一方面,网络已经与人们的学习、工作及生活密不可分;另一方面,网络正朝着多厂商设备、多协议组成、多业务需求的方向渐进,因此实时保障网络的通畅、可靠也就显得极为重要。这除了要依靠网络设备本身和网络架构的可靠性之外,网络管理五大功能之一的故障管理是一个关键环节,它管理的好坏直接影响着网络的运行质量。 由于在过去,网络建设者存在重硬轻软的意识,而且第三方的管理软件也存在价格高、需要更好的技术人员维护管理等因素,就造成了许多网络管理者只是对网络设备、线路和用户进行严格管理,而对于整个网络仅使用设备提供商提供的网管软件的局面。这对于如今的网络来讲是不可靠和不科学的,因此需要一个智能化和简单化的故障管理系统来解决这一问题。 本文以中小型校园网故障管理为目标,研究基于 SNMP 的故障管理系统的组成与实现。论文首先介绍网络管理的五大功能域,详尽阐述 OSI 网络管理框架和体系结构。接着,从简单网络管理协议 SNMP 概况出发,研究 SNMP 的组成部分和运行机制,在总结各种网络管理技术的基础上,阐述了基于 SNMP 的网络故障管理系统的设计原则和开发方法。 论文对基于 SNMP 进行网络故障排除的应用给出了范例。通过介绍故障排除系统的分类、故障排除的方法,使得故障排除流程清晰化、明朗化。接着,结合前人和本人的故障排除经验,将 SNMP MIB 与故障现象联系起来,以实例说明如何利用它们之间的关联关系进行故障排除。 论文还给出了故障管理系统的模块划分与实现。利用分层树状体系结构的思想,设计并部分实现了基于 Web 的四层故障管理系统:利用域管理代理进行数据采集,隔离高层应用与被管设备;使用数据库向上为应用服务提供数据支持,向下接收域管理代理传输的各种数据包;客户端管理界面和浏览器查看页面为不同级别管理员提供了管理功能入口,提高了系统安全性。 本论文对智能化和简单化的网络故障管理系统研究与实现做了初步尝试,结合现代网络管理技术的发展趋势,提出了最终将本系统做成具有高度智能的故障管理系统的实现构想。更多还原

【Abstract】 With the development of computer and communication technology, computer network is rapidly developed too. On one hand, network has already been closely related to people’s study, work and life. On the other hand, it is stepping into a more complex way. So keeping network normal and reliable is very important. This depends not only on the dependability of network device itself and original design, but also on fault management, one of five major functions of network management. The quality of its management influences the running quality of network directly. Therefore, we should observe fault management function at a certain level. How to plan, realize and deploy it to get expected effection in a simple way becomes the main content studied. Because in the past, network builder have the consciousness of emphasising hardware and despising software, and the third-party management software has problems such as high price, need of better technical staff. All of this casuse the situation that many administrators only strictly manage device, circuit and user, but use network management software provided by device manufactory to the whole network. This is not reliable and not scientific to network of today, so we need an intelligent and simplified fault management system to solve this problem. This thesis takes fault management of medium and small-scale campus as its goal, and studies the composition and realization of fault management system based on SNMP. At first, it introduces the five function domains of network management, explains the framework and architecture of OSI network management, which establishes theoretics foundation for later practice. Then, begins with the overview of Simple Network Management Protocol, researches the composition and mechanism of SNMP, summarizes various kinds of network management skills, and shows us the design principle and development approach of the system. Then, this paper gives samples to applications of processing troubleshooting based on SNMP. After the introduction of classification to troubleshooting system and means of fault eliminating, we can clearly know the flow of troubleshooting. Then, combining with my working experience, associates SNMP MIB with fault, and explains how to solve trouble using their relationships with some cases. This thesis also gives the module divisions and practice of the system. According to the hierachical tree-architecture, the web-based four-layer fault management system is designed and partly implemented. There is a data collection subsystem based on domain management agent, which isolates applications from managed device. Database provides data support for the upper applications, and receives all kinds of packets from domain agents. Client management interface and browser interface give differenet level administrator different system entrance, ensuring the security. At last, it implements MIB browser tool using VC++ 6.0, which proved to be correct. Furthermore, this paper does something about intelligent and simplified network fault management system, and still has some deficiency to be improved. Finally, it studies the network management tendencies, and raises a plan of implementing a highly intelligent system in the future.更多还原