基于日志分析的云平台网络故障诊断系统的设计与实现

论文总字数：27719字

摘 要

ABSTRACT ………………………………………………………………………… 3

第一章 绪论 …………………………………………………………………………4

1.1 引言 … ……………………………………………………………………4

1.2网络故障定位技术的研究现状 …………………………………4

1.3 本文的章节安排 …………………………………………………………5

第二章 课题相关技术 …………………………………………………………6

2.1 云计算与openstack ………………………………………………………………6

2.2 数据挖掘技术 ………………………………………………………………9

第三章 网络故障诊断系统的设计 ………………………………………………………11

3.1架构设计 ………………………………………………………………11

3.2模块设计 …………………………………………………………12

3.3.1故障注入 ……………………………………………………………12

3.3.2日志采集 ………………………………………………………13

3.3.3日志预处理 ……………………………………………………14

3.3.4日志挖掘 ………………………………………………………17

3.3.5知识获取 ………………………………………………………20

第四章 系统部署与功能测试 …………………………………………21

4.1系统部署 ……………………………………………………………………21

4.2功能测试 …………………………………………………23

总结 …………………………………………………………………………………26

致谢 …………………………………………………………………………………26

参考文献 ……………………………………………………………………………26

附件 …………………………………………………………………………………28

摘 要

随着云计算的逐渐普及，与云计算的应用也逐渐增多，因此云平台的网络故障诊断具有重要意义。本课题旨在设计并实现一个云平台的网络故障诊断系统，通过故障注入、日志采集、日志预处理、日志挖掘、知识获取等几个步骤，从中发掘有用的知识，进而帮助定位网络故障。本系统的实验环境是基于OpenStack云管理平台的分布式环境，发生网络故障的时候，通过采集、预处理故障日志，与知识库进行对比分析，进而在物理资源层、操作系统层、虚拟机层以及OpenStack层定位网络故障。

基于日志分析的云平台网络故障诊断系统的功能模块包含日志采集模块、日志预处理模块、日志挖掘模块和知识生成模块，用到的技术有正则表达式模式匹配算法、日志压缩与泛化思想、虚拟环境的空间拓扑时变性、聚类边界以及数据挖掘算法——Apriori算法。知识库中存储知识获取模块挖掘出的知识，经过检验与测试，这些挖掘的知识可以有效地帮助定位网络故障。所以本系统对于分布式环境的网络故障定位的具有一定的价值和意义。

关键字：OpenStack、故障定位、日志分析、数据挖掘、关联规则。

ABSTRACT

With the growing popularity of the cloud computing, and the application of cloud computing is gradually increasing, so the cloud platform of network fault diagnosis is of great significance.This subject is to design and implement a cloud platform of network fault diagnosis system, through the fault injection, journal collection, pretreatment, logging, mining, knowledge acquisition, such as several steps, to discover useful knowledge from them to help locate network fault.The experimental environment of this system is based on the distributed environment of the cloud management platform---openstack, when the network fault occurs, through the fault log collection, pretreatment, and the knowledge base, this paper compares and analyzes on the physical resource layer, operating system layer, the virtual machine and openstack layer,then determine the fault occurred in which layer.

Cloud platform network fault diagnosis system based on log analysis including journal acquisition module, pretreatment module, log mining module and knowledge generation module, use of technology has a regular expression pattern matching algorithm, log compressing and generalization, spatial topological time-varying virtual environment, and clustering boundary and the data mining algorithm---Apriori algorithm. Knowledge base excavations of knowledge stored in knowledge acquisition module, after inspection and testing, the knowledge of these mining can effectively help to locate the network failure. Therefore the system for the distributed environment of network fault location has a certain value and significance.

Keywords: OpenStack, Fault Location, Log Analysis, Data Mining, Association Rules.

剩余内容已隐藏，请支付后下载全文，论文总字数：27719字