本文關鍵詞： 網(wǎng)絡模擬 動態(tài)路由 局部路由策略 擁塞預知　出處：《哈爾濱工業(yè)大學》2014年碩士論文　論文類型：學位論文

【摘要】：近年來互聯(lián)網(wǎng)技術(shù)不斷發(fā)展，信息化進程日益推進，互聯(lián)網(wǎng)技術(shù)在改善人們生活方式的同時也帶來了很多問題。一方面，各種網(wǎng)絡安全問題不斷涌現(xiàn)，信息安全事件時有發(fā)生；另一方面，網(wǎng)絡規(guī)模和復雜性的增長對現(xiàn)有的協(xié)議和算法提出了更高的要求，亟需研究人員進行前瞻性的研究。由于網(wǎng)絡模擬在研究網(wǎng)絡行為和網(wǎng)絡協(xié)議方面具有成本低、靈活性強、模擬規(guī)模大等優(yōu)點，因此，得到了越來越廣泛的應用。 在真實的網(wǎng)絡環(huán)境中，路由策略的好壞直接影響到一個網(wǎng)絡的性能，同樣，在網(wǎng)絡模擬環(huán)境中，如何模擬數(shù)據(jù)包在真實網(wǎng)絡環(huán)境中的轉(zhuǎn)發(fā)過程將直接影響模擬的真實性和模擬效率。因此，網(wǎng)絡模擬中路由策略的設計至關重要。 基于上述背景，，本文設計并實現(xiàn)了支持動態(tài)路由的網(wǎng)絡模擬系統(tǒng)，主要完成了如下工作： 第一，研究了當前網(wǎng)絡中廣泛使用的動態(tài)路由協(xié)議，分析了各個協(xié)議的優(yōu)缺點，在NS3上實現(xiàn)了簡化的OSPF協(xié)議模型，主要包括：鏈路狀態(tài)通告的獲取，鏈路狀態(tài)數(shù)據(jù)庫的構(gòu)建，路由選擇算法SPF的實現(xiàn)，路由表的生成以及路由的更新過程。 第二，由于網(wǎng)絡信息量的不斷增加，擁塞已經(jīng)成為制約網(wǎng)絡性能的主要因素，因特網(wǎng)中廣泛使用的OSPF、RIP、IGRP等路由協(xié)議對數(shù)據(jù)包進行路由時總是選擇“最優(yōu)路由”進行轉(zhuǎn)發(fā)，使得鏈路擁塞時其它空閑的鏈路無法得到利用，造成大量網(wǎng)絡資源的浪費。針對這一問題，提出了一種擁塞預知隊列模型，通過周期性的檢測隊列緩沖區(qū)占用率來判斷鏈路的擁塞狀況，當隊列緩沖區(qū)占用率達到一定閾值時，則認為鏈路有發(fā)生擁塞的趨勢。 第三，針對擁塞問題，提出了一種擁塞預知路由算法CPRA，算法采用了局部路由策略的思想：當隊列緩沖區(qū)占用率達到一定閾值時，通過深度優(yōu)先搜索尋找另一條通往目的節(jié)點的路徑，然后將數(shù)據(jù)包沿著該路徑轉(zhuǎn)發(fā)。在NS3上的仿真結(jié)果表明，相對于OSPF，使用CPRA能夠顯著提高網(wǎng)絡的性能，其中，網(wǎng)絡吞吐率提高了約50%，端到端平均時延減少了0.3s。
[Abstract]:In recent years, with the continuous development of Internet technology, the process of information technology is advancing day by day. Internet technology not only improves people's way of life, but also brings a lot of problems. On the one hand, a variety of network security problems continue to emerge. Information security incidents occur from time to time; On the other hand, the growth of network size and complexity puts forward higher requirements for existing protocols and algorithms. Because network simulation has the advantages of low cost, strong flexibility and large simulation scale in studying network behavior and network protocol. Has been more and more widely used. In the real network environment, the routing strategy directly affects the performance of a network, similarly, in a network simulation environment. How to simulate the forwarding process of packets in real network environment will directly affect the authenticity and efficiency of simulation. Therefore, the design of routing strategy in network simulation is very important. Based on the above background, a network simulation system supporting dynamic routing is designed and implemented in this paper. The main work is as follows: First, the dynamic routing protocols widely used in the current network are studied, the advantages and disadvantages of each protocol are analyzed, and the simplified OSPF protocol model is implemented on NS3. It mainly includes: the acquisition of link state notification, the construction of link state database, the implementation of routing algorithm SPF, the generation of routing table and the updating process of routing. Secondly, due to the increasing amount of network information, congestion has become the main factor that restricts the network performance. OSPFN RIP, which is widely used in the Internet, is widely used. IGRP and other routing protocols always select the "optimal route" to forward the packets, which makes the other idle links can not be utilized when the link is congested. In order to solve this problem, a congestion prediction queue model is proposed to determine the congestion status of the link by periodically detecting queue buffer occupancy. When queue buffer occupancy reaches a certain threshold, the link is considered to be congested. Thirdly, a congestion predictive routing algorithm, CPRA-based, is proposed, which adopts the idea of local routing strategy: when queue buffer occupancy reaches a certain threshold. Another path to the destination node is found by depth first search, and then the packet is forwarded along the path. The simulation results on NS3 show that it is relative to OSPF. The performance of the network can be significantly improved by using CPRA. The throughput of the network increases by about 50 and the average end-to-end delay decreases by 0.3 s.
【學位授予單位】：哈爾濱工業(yè)大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TP393.08

【參考文獻】

相關期刊論文 前5條

1 李姍姍;廖湘科;朱培棟;肖儂;;傳感器網(wǎng)絡中一種擁塞避免、檢測與緩解策略[J];計算機研究與發(fā)展;2007年08期

2 王曉鋒;關鷺;劉淵;;基于拓撲抽象的高性能網(wǎng)絡模擬方法[J];計算機工程與應用;2012年29期

3 尹偉力;;無尺度網(wǎng)絡模型與通訊工具的發(fā)展趨勢[J];計算機與信息技術(shù);2006年12期

4 張兆心;杜躍進;王克;丁振全;郝志宇;;基于焦點折疊的網(wǎng)絡模擬拓撲抽象模型[J];通信學報;2012年07期

5 陳巖;董淑福;蔣磊;;OPNET網(wǎng)絡仿真技術(shù)及其應用研究[J];計算機技術(shù)與發(fā)展;2009年02期

相關博士學位論文 前1條

1 毛鵬軒;下一代網(wǎng)絡擁塞控制關鍵算法的研究[D];北京交通大學;2013年

本文編號：1483299