本文選題：擁塞控制 + XCP協(xié)議�。� 參考：《山東師范大學》2014年碩士論文

【摘要】：計算機技術的飛速發(fā)展以及互聯(lián)網(wǎng)時代的到來，翻開了人們生活方式的新篇章�；ヂ�(lián)網(wǎng)世界逐漸發(fā)展壯大，人們的生活需求也在不斷的變化。進入大數(shù)據(jù)時代，高速數(shù)據(jù)（如視頻流等）傳輸、特定數(shù)據(jù)傳輸服務成為主流，網(wǎng)絡環(huán)境也因此發(fā)生了相應的改變。隨著新的網(wǎng)絡環(huán)境的出現(xiàn)，以及網(wǎng)絡世界呈現(xiàn)出的高帶寬時延積特性，擁塞控制再次成為人們的重要關注點。而傳統(tǒng)的TCP擁塞控制機制在網(wǎng)絡環(huán)境變化的驚濤駭浪中顯示出很多性能上的不足。 針對出現(xiàn)的問題，很多中外學者對新環(huán)境下的擁塞控制機制進行了大量的研究，一部分學者進行了基于源端擁塞控制算法的研究，在原TCP協(xié)議上進行改進研究，取得了一定的成果。而另一部分學者則提出了基于鏈路的控制方法，在路由器中加入擁塞指示的反饋，將中間通信網(wǎng)絡引入到整個擁塞控制機制中，使得數(shù)據(jù)發(fā)送端能夠顯式地接收當前的網(wǎng)絡狀況信息。經(jīng)過大量實驗以及理論分析，顯式擁塞指示愈來愈體現(xiàn)出其優(yōu)越性。而其中的XCP協(xié)議，以其表現(xiàn)出的高性能而獲得學者的青睞，成為新網(wǎng)絡環(huán)境下進行網(wǎng)絡擁塞控制的新趨勢。 本文對擁塞控制的基本原理進行了講述，理解傳統(tǒng)網(wǎng)絡中TCP協(xié)議的工作原理，并根據(jù)當前網(wǎng)絡環(huán)境的變化對TCP協(xié)議進行了重新評價。隨著高帶寬時延積網(wǎng)絡的出現(xiàn)與發(fā)展，TCP協(xié)議不能保持系統(tǒng)工作的高效率性，從而無法快速高效地利用網(wǎng)絡資源。同時，由于多種應用需求的出現(xiàn)，數(shù)據(jù)流的公平性要求逐漸升高，而傳統(tǒng)TCP協(xié)議無法很好地實現(xiàn)對每條數(shù)據(jù)流的公平分配，這成為TCP協(xié)議在新網(wǎng)絡環(huán)境中發(fā)展的瓶頸。除此之外，高帶寬時延積網(wǎng)絡中的TCP協(xié)議無法保證系統(tǒng)的穩(wěn)定性，使得網(wǎng)絡容易陷于抖動之中，很容易導致網(wǎng)絡擁塞，甚至崩潰。傳統(tǒng)的擁塞控制在新的網(wǎng)絡環(huán)境中遇到了挑戰(zhàn)，而顯式擁塞指示的出現(xiàn)讓學者們找到了新的研究方向。本文對顯式擁塞指示進行了詳細講解，以便于理解XCP協(xié)議的發(fā)展背景。同時，本文還對系統(tǒng)的穩(wěn)定性判定定理進行了一定的講解，以便于對XCP協(xié)議系統(tǒng)的穩(wěn)定性分析。 本文詳細描述了XCP協(xié)議的工作原理，從宏觀的角度整體掌握了XCP協(xié)議的工作流程。支持XCP協(xié)議的數(shù)據(jù)發(fā)送端將帶有擁塞信息的數(shù)據(jù)包發(fā)送到網(wǎng)絡中，中間通信網(wǎng)絡的路由器接收到數(shù)據(jù)包之后，根據(jù)當前的網(wǎng)絡情況修改數(shù)據(jù)包中的擁塞信息，數(shù)據(jù)包到達目的端，目的端將數(shù)據(jù)包中最終的擁塞信息以反饋信息的方式發(fā)送給發(fā)送端，發(fā)送端根據(jù)反饋信息了解到當前的網(wǎng)絡狀況，并以此來調整下一周期的發(fā)送速率。要完成這一系列的控制工作，需要在數(shù)據(jù)包中加入存放擁塞信息的擁塞控制頭，路由器在估算出當前的網(wǎng)絡狀況后修改該頭結構內的信息。作為XCP協(xié)議的關鍵參與部分，XCP路由器的工作主要分為保證效率性和公平性兩部分，分別由效率控制器和公平控制器來完成。 本文重點從XCP協(xié)議中資源利用的角度進行了分析，通過了解資源計算對資源利用的重要性，分別從參數(shù)分析以及帶寬補償兩方面進行資源分析方法的講解。效率控制器中的參數(shù)變化對系統(tǒng)的性能產(chǎn)生了影響，通過調整參數(shù)的選擇方式可以進一步調高算法的資源利用率，并保證整個系統(tǒng)的穩(wěn)定性以及收斂性。帶寬補償是另一種改善資源計算的方式。通過對控制器的重新調整，在原控制器的基礎上進行改進，從而提升整個機制的性能。除此之外，本文提出了一種新的計算網(wǎng)絡資源的方法，，除每個控制周期計算的可分配資源外，增加一部分帶寬資源，減輕了原算法保守的特性。同時，新算法將帶寬重洗考慮在內，更加符合實際情況。實驗仿真顯示，新算法能夠更充分的利用網(wǎng)絡資源，并能夠保證系統(tǒng)在閾值范圍內的穩(wěn)定性。 本文在對XCP協(xié)議進行分析、仿真以及改進之后，從控制論的角度對XCP協(xié)議進行了理論分析。利用常微分方程描述系統(tǒng)，隨后利用李亞普諾夫方法，分別對XCP系統(tǒng)的一般情況以及考慮實際時延的情況進行了分析，驗證系統(tǒng)的穩(wěn)定性條件。
[Abstract]:The rapid development of computer technology and the advent of the Internet era have opened a new chapter of people's life style. The Internet world is growing and growing, and the needs of people are constantly changing. Entering the era of big data, high speed data (such as video stream) transmission, specific data transmission services become the mainstream, and the network environment is also issued. With the emergence of the new network environment and the characteristics of high bandwidth delay product in the network world, congestion control has become an important concern again. The traditional TCP congestion control mechanism shows a lot of inadequacy in the wave of network environment.
In view of the problem, many Chinese and foreign scholars have done a lot of research on the congestion control mechanism under the new environment. Some scholars have done the research on the source based congestion control algorithm, improved the original TCP protocol and achieved some results. While another part of the scholars proposed a link based control method, in the routing. With the feedback of congestion indication, the intermediate communication network is introduced into the whole congestion control mechanism, which enables the data transmitter to receive the current network information explicitly. After a large number of experiments and theoretical analysis, the explicit congestion indication shows its superiority more and more. The XCP protocol shows high performance. It has become the new trend of network congestion control under the new network environment.
The basic principle of congestion control is described in this paper, the working principle of TCP protocol in the traditional network is understood, and the TCP protocol is reevaluated according to the changes of current network environment. With the emergence and development of the high bandwidth delay product network, the TCP protocol can not keep the efficiency and efficiency of the system, so that it can not be used quickly and efficiently. At the same time, due to the emergence of a variety of application requirements, the demand for the fairness of data flow is increasing gradually, and the traditional TCP protocol can not achieve a fair distribution of each data stream. This has become the bottleneck of the development of the TCP protocol in the new network environment. In addition, the TCP protocol in the high bandwidth delay product network can not guarantee the stability of the system. Qualitatively, the network is prone to jitter, which can easily lead to network congestion and even collapse. Traditional congestion control has met challenges in the new network environment, and the emergence of explicit congestion indication has made scholars find new research directions. This paper explains the explicit congestion indication in detail in order to understand the development of XCP protocol. Background, at the same time, we explain the stability theorem of the system in order to analyze the stability of the XCP protocol system.
In this paper, the working principle of XCP protocol is described in detail, and the workflow of the XCP protocol is mastered from a macro point of view. The data packets with congestion information are sent to the network by the data sending end of the XCP protocol. After the router receives the packet, the congestion in the packet is modified according to the current network situation. Information, the packet arrives at the destination, the destination sends the final congestion information in the packet to the sending end in the way of feedback information. The sender understands the current network status according to the feedback information and adjusts the sending rate of the next cycle. Congestion control head of plug information, the router modifies the information in the header structure after estimating the current network condition. As the key part of the XCP protocol, the work of XCP router is divided into two parts: efficiency and fairness, which are completed by efficiency controller and fair controller respectively.
This paper focuses on the analysis of resource utilization in the XCP protocol. Through understanding the importance of resource utilization in resource computing, the paper explains the resource analysis methods from two aspects of parameter analysis and bandwidth compensation. The parameters change in the efficiency controller affects the performance of the system and adjusts the selection of parameters. It can further improve the resource utilization of the algorithm, and ensure the stability and convergence of the whole system. Bandwidth compensation is another way to improve the computing of resources. By the adjustment of the controller, the performance of the whole mechanism is improved on the basis of the original controller. In addition, a new method is proposed. The method of computing network resources, in addition to the allocated resources calculated in each control cycle, increases a part of the bandwidth resources and reduces the conservative characteristics of the original algorithm. At the same time, the new algorithm takes the bandwidth reshuffling into consideration and is more in line with the actual situation. The experimental simulation shows that the new algorithm can make full use of the network resources and can ensure the system in the threshold. Stability within the range of values.
In this paper, after the analysis, simulation and improvement of the XCP protocol, the XCP protocol is theoretically analyzed from the angle of control theory. The system is described by using ordinary differential equations. Then, the general situation of the XCP system and the situation of the actual delay are analyzed by Lyapunov method, and the stability conditions of the system are verified.

【學位授予單位】：山東師范大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TP393.04

【參考文獻】

相關期刊論文 前9條

1 周海瑞;戴冠中;郭達偉;張慧翔;;基于PII控制器的XCP帶寬補償算法[J];電子學報;2009年09期

2 張慧翔;戴冠中;姚磊;潘文平;;一種提高XCP協(xié)議在大RTT差異環(huán)境下的魯棒性方法[J];計算機科學;2008年07期

3 張慧翔;戴冠中;姚磊;潘文平;;一種增強XCP協(xié)議在多瓶頸網(wǎng)絡拓撲環(huán)境下性能的方法[J];計算機科學;2008年12期

4 張福杰;潘理;李建華;;大帶寬時延積網(wǎng)絡中TCP,HighSpeed TCP及XCP性能比較[J];計算機工程;2006年02期

5 陳元琰;吳東;陶小梅;;XCP中a參數(shù)的自適應動態(tài)調整算法[J];計算機工程;2006年22期

6 黃小猛;林闖;任豐源;;高速傳輸協(xié)議研究進展[J];計算機學報;2006年11期

7 盧利瓊;陳元琰;吳東;羅曉曙;閆友彪;;高帶寬時延乘積網(wǎng)絡中的擁塞控制端算法研究[J];計算機應用研究;2006年03期

8 任豐原,林闖,任勇,山秀明;大時滯網(wǎng)絡中的擁塞控制算法[J];軟件學報;2003年03期

9 王德鎖,舒勤;TCP擁塞控制機制在高速網(wǎng)絡中的局限性[J];中國測試技術;2004年02期

本文編號：1799654