本文選題：分布式 + 網(wǎng)站監(jiān)測�。� 參考：《北京化工大學》2014年碩士論文

【摘要】：互聯(lián)網(wǎng)技術自發(fā)明以來,走過了40個年頭,其發(fā)展迅速性和規(guī)模龐大性使其應用廣泛,遍及各個行業(yè)。互聯(lián)網(wǎng)最大的應用之一網(wǎng)站,已經(jīng)滲透了我國各個角落,很多網(wǎng)站存在可訪問性差、用戶對網(wǎng)站的滿意度不高,以及網(wǎng)站已經(jīng)宕機而無人知曉等問題,不能及時的發(fā)現(xiàn)網(wǎng)站的運行狀況和可用性問題。 前人對網(wǎng)站監(jiān)測只是采用單點測量數(shù)據(jù)判斷網(wǎng)站的運行狀況,最終結果準確率較低,并且缺少網(wǎng)站預測的功能。針對前人對網(wǎng)站監(jiān)測研究的不足,為保證網(wǎng)站的安全穩(wěn)定運行,通過分布式監(jiān)測來提高網(wǎng)站預測準確率。論文首先介紹了網(wǎng)站的性能指標參數(shù),構建分布式網(wǎng)站仿真工且OPNET、信息�；南嚓P知識。其次介紹了網(wǎng)站監(jiān)測分類和網(wǎng)站性能狀態(tài)的評測標準,參考國內(nèi)外研究成果,設計了一個分布式網(wǎng)站監(jiān)測系統(tǒng),采集分布式網(wǎng)站監(jiān)測系統(tǒng)的仿真數(shù)據(jù)進行了Web性能參數(shù)之間關系的分析。最后,采用單點監(jiān)測和分布式監(jiān)測兩種方式對實際網(wǎng)站進行監(jiān)測。對網(wǎng)站的狀態(tài)分類采用隨機驗證、交叉選擇、遺傳算法和粒子群算法進行優(yōu)化和對比說明。對網(wǎng)站狀態(tài)預測采用基于支持向量機的信息粒化的方式進行預測,對一段時間內(nèi)響應時間的最小值預測的相對準確率可以達到95%,平均值相對準確率達到96.2%,最大值的相對準確率 達到87%,可較準確預測出網(wǎng)站響應時間的范圍,保證網(wǎng)站健康、穩(wěn)定的運行,改善網(wǎng)站的可用性和用戶滿意度。課題的成功預測對日益嚴重的網(wǎng)絡安全問題具有重要的研究意義和應用價值。
[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.
【學位授予單位】：北京化工大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TP393.092

【參考文獻】

相關期刊論文 前10條

1 王有為,汪定偉;基于網(wǎng)頁可達性和平均載入時間的網(wǎng)站評估方法[J];東北大學學報;2002年10期

2 王宏偉;楊雨;張智華;;網(wǎng)絡時延預測研究[J];電腦知識與技術;2011年10期

3 夏中林;;校園網(wǎng)網(wǎng)站監(jiān)測系統(tǒng)的設計與實現(xiàn)[J];電腦知識與技術;2012年08期

4 姜明;吳春明;張e，

本文編號：1833050