中国科学院软件研究所机构知识库
Advanced  
ISCAS OpenIR  > 中科院软件所  > 中科院软件所
题名:
以太网交换机中服务质量保证机制的研究
作者: 王显雷
答辩日期: 2007-06-05
授予单位: 中国科学院软件研究所
授予地点: 软件研究所
学位: 博士
关键词: 服务质量(QoS) ; 以太网交换机 ; 二层组播QoS最优生成树 ; 最优根桥逼近查找算法 ; 可控SVL(CSVL) ; 非对称VLAN ; 缓冲区容量 ; 队列调度算法
其他题名: Research on QoS Mechanisms in Ethernet Switch
摘要: 随着计算机网络承载业务的多样化,如何提高网络服务质量已经成为网络研究的热点。本文着眼于改善二层域中各种业务的服务质量,围绕以太网交换机,研究了二层域中各种服务质量保证机制。本文的主要贡献和创新点包括: 第一,从二层组播接收者的角度出发,提出了二层组播QoS最优生成树的概念,从理论上证明了组播源位于最优生成树的根网桥上时,组播能达到最优的服务质量。而且,最优生成树对于经过根网桥的单播也能达到最优的服务质量。本文给出构建最优生成树的最优根网桥逼近查找算法,通过对比实验验证了该算法的有效性、可靠性和可扩展性。最优根网桥逼近查找算法可以作为生成树算法的补充,使得生成树协议在选择根网桥时,能考虑到其对组播服务质量的影响,这能有效地改善具有一对多特性组播数据的服务质量。 第二,非对称VLAN基于交换机的不同MAC地址学习及转发模型有着不同的实现。基于本文提出的可控SVL(Controllable Shared VLAN Learning,简称CSVL)MAC地址学习及转发模型的非对称VLAN,在不增加额外MAC地址表项的情况下,能够实现非对称VLAN的功能,并具有良好的适应性和可扩展性。同时能有效降低非对称VLAN管理协议的复杂度,减少对CPU时间和交换芯片处理时间的占用。为实现不同用户隔离和共享资源双重目的的非对称VLAN,能够以二层交换的速度实现三层路由的功能,能大大提高交换速度,改善服务质量。 第三,本文明确了缓冲区研究中的几个关键问题,分析TCP拥塞控制算法的理想化模型,得出了TCP最大发送窗口的大小是由“带宽-时延积”和缓冲区容量共同决定的结论,给出定量关系。从而使得缓冲区容量和丢包率、吞吐量及排队时延等服务质量衡量标准建立了间接的量化关系,得出了缓冲区容量设置的规约化公式。仿真结果能很好地吻合本文得出的量化关系公式。缓冲区容量对网络服务质量具有显著影响。一般来说,随着缓冲区增大,输出链路的带宽利用率升高,丢包率降低,同时,数据包的最大排队时延也会增大。 第四,本文提出了输出链路带宽时隙(slot)化的概念,分析了目前队列机制的漏桶本质,提出了基于令牌桶的队列调度算法,将时隙化的带宽作为令牌,缓存一定量输出链路带宽的使用权限,获得了更好的公平性和适应能力,能适应流量的突发,降低丢包率。队列调度算法按照一定的份额和顺序控制多个业务流共享输出链路带宽,将直接影响到业务流的带宽分配、丢包率、时延和时延抖动等服务质量特性。队列调度算法是交换机实施区分服务的核心机制之一。
英文摘要: With the diversification of traffic that computer network transports, how to improve the quality of service (QoS) is becoming the hot problem in the field of computer network. This dissertation focuses on how to improve the QoS in layer 2 domain, and researches into a variety of QoS mechanism in layer 2 domain surrounding Ethernet switch. The main contributions of this dissertation as follows: Firstly,Recently, more and more QoS-sensitive multicast traffic comes into Ethernet, which requires higher quality of service(QoS) than traditional data traffic. As a result, the current spanning tree protocol(STP) can not provide satisfactory QoS for the new type traffic. Because multicast traffic has inherent P2MP characteristic, where the root bridge of STP is located greatly affects its QoS in layer 2. However, the current STP just selects the bridge that owns the smallest bridge ID as its root bridge, leaving QoS out of consideration. In this paper, from the point of view of multicast receivers, we bring forward the L2 multicast QoS optimized spanning tree concept, and prove that if multicast source is mobile, when multicast source is located at the root bridge of optimized spanning tree, multicast traffic gains optimized QoS. Otherwise, the bridge that multicast source is located at is selected as the root bridge of STP in order to get best multicast QoS. In addition, we show unicast traffic, which shuttles through the root bridge of optimized spanning tree, gains optimized QoS too. Finally, we propose an approximating lookup algorithm of the optimized root bridge as the complement of the spanning tree algorithm and verify the validity and reliability of this algorithm in our experiment. Secondly,In virtual bridged local area networks, asymmetric VLAN was brought forward in order to segregate users in different VLAN, at the same time, all users regardless of their VLAN membership can access shared resources in server VLAN. we first analyze the shortage of SVL-based Asymmetric VLAN and IVL-based Asymmetric VLAN. Then, we provide a novel Controllable Shared VLAN Learning(CSVL) and CSVL-based Asymmetric VLAN. At last, we verify CSVL-based Asymmetric VLAN can achieve the function of Asymmetric VLAN perfectly without adding a great amount of extra MAC address entries in FDB by our experiment. CSVL-based Asymmetric VLAN can also reduce the complexity of Asymmetric VLAN management protocol and cut down the usage time of CPU and switch chip. Thirdly,All Ethernet switch contain buffers to hold packets during times of congestion. Buffers cause queueing delay and delay-variance; when they overflow they cause packet loss, and when they underflow they can degrade throughput.In the paper, we clarify several key issues regarding buffer sizing. After analyzing carefully the idealized model of the dynamics of TCP’s congestion control algotithms, we draw a conclusion: maximum TCP send windows size is determined by buffer size and the link’s bandwidth-delay product together, and give the quantitative relation formula between them. Indirectly, we get the the quantitative relation of buffer sizing with loss probability, throughout and queuing delay. Using simulation, we show simulation results can perfectly match the quantitative relation formula we deduce in theory. Finally,In the paper, token bucket-based scheduling algorithm was brought forward. It can solve “starvation” problem of strict priority queue scheduling algorithm.and adapt burst of traffic decreasing packets loss. Scheduling algorithm is one of the core mechanics to implement differentiated services.
语种: 中文
内容类型: 学位论文
URI标识: http://ir.iscas.ac.cn/handle/311060/7156
Appears in Collections:中科院软件所

Files in This Item:
File Name/ File Size Content Type Version Access License
10001_200318015003121王显雷_paper.pdf(2231KB)----限制开放-- 联系获取全文

Recommended Citation:
王显雷. 以太网交换机中服务质量保证机制的研究[D]. 软件研究所. 中国科学院软件研究所. 2007-06-05.
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[王显雷]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[王显雷]‘s Articles
Related Copyright Policies
Null
Social Bookmarking
Add to CiteULike Add to Connotea Add to Del.icio.us Add to Digg Add to Reddit
所有评论 (0)
暂无评论
 
评注功能仅针对注册用户开放,请您登录
您对该条目有什么异议,请填写以下表单,管理员会尽快联系您。
内 容:
Email:  *
单位:
验证码:   刷新
您在IR的使用过程中有什么好的想法或者建议可以反馈给我们。
标 题:
 *
内 容:
Email:  *
验证码:   刷新

Items in IR are protected by copyright, with all rights reserved, unless otherwise indicated.

 

 

Valid XHTML 1.0!
Copyright © 2007-2017  中国科学院软件研究所 - Feedback
Powered by CSpace