發(fā)布時(shí)間：2018-06-15 09:56

  本文選題：網(wǎng)絡(luò)流量監(jiān)控 + 時(shí)間序列��； 參考：《天津理工大學(xué)》2014年碩士論文

【摘要】：計(jì)算機(jī)網(wǎng)絡(luò)的發(fā)展，使得信息的交流和資源的共享更加便捷。為了教師教學(xué)和學(xué)生學(xué)習(xí)的方便，校園網(wǎng)帶寬逐年擴(kuò)大，訪問(wèn)的速度也得到了很大的提高。但是，目前校園網(wǎng)帶寬的有效利用率并不高，大部分帶寬用于游戲、視頻、即時(shí)通信以及P2P應(yīng)用。由于這些與工作關(guān)系不大的應(yīng)用，搶占、消耗大量的網(wǎng)絡(luò)帶寬資源，降低了正常通信的質(zhì)量，影響了正常的工作和學(xué)習(xí)。因此，如何準(zhǔn)確地掌控網(wǎng)絡(luò)運(yùn)行狀況，分析網(wǎng)絡(luò)應(yīng)用服務(wù)，進(jìn)行有效地網(wǎng)絡(luò)監(jiān)控，提高網(wǎng)絡(luò)帶寬有效利用率，成為亟需解決的問(wèn)題，具有重要的研究意義。 本文旨在通過(guò)研究校園網(wǎng)流量基本特征，構(gòu)建流量預(yù)測(cè)模型，并利用流量的歷史數(shù)據(jù)預(yù)測(cè)未來(lái)流量變化趨勢(shì)。運(yùn)用應(yīng)用層協(xié)議分析技術(shù)對(duì)不同種類的網(wǎng)絡(luò)應(yīng)用服務(wù)進(jìn)行統(tǒng)計(jì)分析，管理員根據(jù)預(yù)測(cè)和分析結(jié)果進(jìn)行網(wǎng)絡(luò)運(yùn)行控制，對(duì)P2P應(yīng)用、游戲以及視頻等應(yīng)用服務(wù)進(jìn)行帶寬限制。 本文的主要研究?jī)?nèi)容是在認(rèn)真分析了校園網(wǎng)絡(luò)流量自相似性以及時(shí)間序列分析技術(shù)的基本概念、理論方法和建模步驟的基礎(chǔ)上，進(jìn)一步深入研究了自回歸移動(dòng)平均模型(ARMA)、季節(jié)性自回歸移動(dòng)平均模型(SARMA)和廣義自回歸條件異方差模型(GARCH)的建模過(guò)程；其次，對(duì)網(wǎng)絡(luò)流量監(jiān)控與分析技術(shù)進(jìn)行了研究，包括應(yīng)用層協(xié)議分析技術(shù)、多模式狀態(tài)機(jī)匹配技術(shù)和簡(jiǎn)單網(wǎng)絡(luò)管理協(xié)議(SNMP)。通過(guò)實(shí)驗(yàn)效果對(duì)比，最終提出了一種運(yùn)用SARMA模型對(duì)網(wǎng)絡(luò)流量歷史數(shù)據(jù)進(jìn)行建模，并利用GARCH模型對(duì)其殘差進(jìn)行修正的網(wǎng)絡(luò)流量預(yù)測(cè)模型建模方法。 論文運(yùn)用時(shí)間序列分析技術(shù)建立網(wǎng)絡(luò)流量預(yù)測(cè)模型，預(yù)測(cè)網(wǎng)絡(luò)流量下一個(gè)時(shí)間段內(nèi)的變化趨勢(shì)。本文還采用端口技術(shù)、深度包檢測(cè)技術(shù)(DPI)、深度流檢測(cè)技術(shù)(DFI)進(jìn)行應(yīng)用層協(xié)議分析，在協(xié)議分析過(guò)程中采用多模式狀態(tài)機(jī)匹配算法，進(jìn)行特征匹配，，提高了匹配的速度和準(zhǔn)確度，最后通過(guò)SNMP對(duì)網(wǎng)絡(luò)設(shè)備進(jìn)行設(shè)置，從而達(dá)到對(duì)某一IP或應(yīng)用協(xié)議進(jìn)行控制的目的。 結(jié)合使用以上各項(xiàng)技術(shù)，設(shè)計(jì)并實(shí)現(xiàn)了網(wǎng)絡(luò)流量預(yù)測(cè)與監(jiān)控系統(tǒng)。系統(tǒng)基于B/S架構(gòu)，前臺(tái)頁(yè)面使用ExtJS4.0進(jìn)行頁(yè)面渲染，后臺(tái)在SSH(Struts2+Spring+Hibernate)框架下進(jìn)行MVC模式開發(fā)，前后臺(tái)數(shù)據(jù)交互過(guò)程使用JSON傳輸，整個(gè)管理系統(tǒng)在Web方式下以Desktop樣式呈現(xiàn)。論文介紹了網(wǎng)絡(luò)流量預(yù)測(cè)與監(jiān)控系統(tǒng)的設(shè)計(jì)方法，并對(duì)各個(gè)功能模塊的實(shí)現(xiàn)方法進(jìn)行了詳盡地闡述，最后對(duì)系統(tǒng)部署以及性能分析進(jìn)行了描述。
[Abstract]:With the development of computer network, the exchange of information and the sharing of resources are more convenient. For the convenience of teachers' teaching and students' study, the bandwidth of campus network has been enlarged year by year, and the speed of access has been greatly improved. However, the current campus network bandwidth utilization is not high, most of the bandwidth for games, video, instant messaging and P2P applications. Because of these little work related applications, preemption, consumption of a large number of network bandwidth resources, reduce the quality of normal communications, affect the normal work and learning. Therefore, how to accurately control the operation of the network, analyze the network application services, effectively monitor the network, improve the effective utilization of network bandwidth, become an urgent problem, which has an important research significance. The purpose of this paper is to build a traffic forecasting model by studying the basic characteristics of campus network traffic, and to predict the trend of traffic change in the future by using the historical data of traffic. The application layer protocol analysis technology is used to analyze the different kinds of network application services. The administrator controls the network operation according to the prediction and analysis results, and restricts the bandwidth of P2P applications, games, video and other application services. The main research content of this paper is based on the analysis of the basic concepts, theoretical methods and modeling steps of the campus network traffic self-similarity and time series analysis technology. The modeling process of autoregressive moving average model, seasonal autoregressive moving average model and generalized autoregressive conditional heteroscedasticity model (GARCH) are further studied. Secondly, the network traffic monitoring and analysis techniques are studied. It includes application layer protocol analysis technology, multi-mode state machine matching technology and simple network management protocol (SNMP). Through the comparison of experimental results, a modeling method of network traffic prediction model using SARMA model is proposed, and the residual error is modified by GARCH model. In this paper, the time series analysis technique is used to establish a network traffic prediction model to predict the trend of network traffic in the next time period. This paper also uses port technology, depth packet detection technology and depth flow detection technology to analyze the protocol in application layer. In the process of protocol analysis, multi-mode state machine matching algorithm is used to match features, and the speed and accuracy of matching are improved. Finally, the network device is set up by SNMP to control a certain IP or application protocol. A network traffic forecasting and monitoring system is designed and implemented by using the above technologies. The system is based on the structure of B / S, the front page is rendered by Ext JS4.0, the background is developed by MVC pattern under the framework of SSH / Struts2 Spring hibernate, the process of front and back data exchange is transmitted by JSON, and the whole management system is presented as Desktop style in Web mode. This paper introduces the design method of network traffic prediction and monitoring system, and describes the implementation of each functional module in detail. Finally, the system deployment and performance analysis are described.
【學(xué)位授予單位】：天津理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：TP393.06;O211.61

【參考文獻(xiàn)】

相關(guān)期刊論文 前6條

1 王珊;陳健;黃志根;;基于狀態(tài)機(jī)的應(yīng)用層協(xié)議識(shí)別和內(nèi)容分析[J];電子測(cè)量技術(shù);2011年12期

2 陳亮;龔儉;徐選;;基于特征串的應(yīng)用層協(xié)議識(shí)別[J];計(jì)算機(jī)工程與應(yīng)用;2006年24期

3 王杰;石成輝;;基于正則表達(dá)式的動(dòng)態(tài)應(yīng)用層協(xié)議識(shí)別方案[J];計(jì)算機(jī)工程與應(yīng)用;2010年18期

4 陳曉天;張順頤;田婷婷;;基于BP神經(jīng)網(wǎng)絡(luò)的IP網(wǎng)絡(luò)流量預(yù)測(cè)[J];南京郵電大學(xué)學(xué)報(bào)(自然科學(xué)版);2010年02期

5 劉宏偉,楊孝宗,曲峰,趙金華;非齊次泊松過(guò)程類軟件可靠性增長(zhǎng)模型[J];同濟(jì)大學(xué)學(xué)報(bào)(自然科學(xué)版);2004年08期

6 孫宏旭;邢薇;陶林;;基于有限狀態(tài)機(jī)的模型轉(zhuǎn)換方法的研究[J];計(jì)算機(jī)技術(shù)與發(fā)展;2012年02期

相關(guān)博士學(xué)位論文 前1條

1 賈煒;計(jì)算機(jī)網(wǎng)絡(luò)脆弱性評(píng)估方法研究[D];中國(guó)科學(xué)技術(shù)大學(xué);2012年

本文編號(hào)：2021618