分布式网站监测及性能分析研究

发布时间：2018-05-02 07:59

本文选题：分布式 + 网站监测　；参考：《北京化工大学》2014年硕士论文

【摘要】：互联网技术自发明以来,走过了40个年头,其发展迅速性和规模庞大性使其应用广泛,遍及各个行业。互联网最大的应用之一网站,已经渗透了我国各个角落,很多网站存在可访问性差、用户对网站的满意度不高,以及网站已经宕机而无人知晓等问题,不能及时的发现网站的运行状况和可用性问题。前人对网站监测只是采用单点测量数据判断网站的运行状况,最终结果准确率较低,并且缺少网站预测的功能。针对前人对网站监测研究的不足,为保证网站的安全稳定运行,通过分布式监测来提高网站预测准确率。论文首先介绍了网站的性能指标参数,构建分布式网站仿真工且OPNET、信息粒化的相关知识。其次介绍了网站监测分类和网站性能状态的评测标准,参考国内外研究成果,设计了一个分布式网站监测系统,采集分布式网站监测系统的仿真数据进行了Web性能参数之间关系的分析。最后,采用单点监测和分布式监测两种方式对实际网站进行监测。对网站的状态分类采用随机验证、交叉选择、遗传算法和粒子群算法进行优化和对比说明。对网站状态预测采用基于支持向量机的信息粒化的方式进行预测,对一段时间内响应时间的最小值预测的相对准确率可以达到95%,平均值相对准确率达到96.2%,最大值的相对准确率达到87%,可较准确预测出网站响应时间的范围,保证网站健康、稳定的运行,改善网站的可用性和用户满意度。课题的成功预测对日益严重的网络安全问题具有重要的研究意义和应用价值。
[Abstract]:Internet technology has been developed for 40 years since its invention. Its rapid development and large scale make it widely used in various industries. One of the largest applications of the Internet, websites have penetrated every corner of our country. Many websites have problems such as poor accessibility, low user satisfaction with the website, and the fact that the website has gone down without anyone knowing it. Unable to find out the running status and usability of the website in time. The former people only use single point measurement data to judge the running condition of the website, and the accuracy of the final result is low, and the function of website prediction is lacking. In order to ensure the safe and stable operation of the website, distributed monitoring is used to improve the accuracy of website prediction. Firstly, the paper introduces the performance parameters of the website, constructs the distributed website simulator and OPNETand the relevant knowledge of information granulation. Secondly, this paper introduces the classification of website monitoring and the evaluation standard of website performance status, and designs a distributed website monitoring system with reference to domestic and foreign research results. The simulation data of distributed website monitoring system are collected and the relationship between Web performance parameters is analyzed. Finally, single-point monitoring and distributed monitoring are used to monitor the actual website. This paper uses random verification, cross selection, genetic algorithm and particle swarm optimization to optimize and explain the status classification of the website. The information granulation method based on support vector machine (SVM) is used to predict the state of the website. The relative accuracy of predicting the minimum response time over a period of time can reach 95%, the relative accuracy of average value reaches 96.2and the relative accuracy of maximum value reaches 96.2%. It can accurately predict the response time range of the website, ensure the healthy and stable operation of the website, improve the usability and user satisfaction of the website. The successful prediction of this topic has important significance and application value to the increasingly serious network security problems.
【学位授予单位】：北京化工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.092

【参考文献】