基于IPv6的校园网技术升级及性能测量研究
|
摘要
以IPv4协议为基础的互联网已经走过了四十余年的发展历程，逐步暴露出诸如地址短缺、网络安全性等突出问题。而IPv6与IPv4相比，以其更大、更快、更安全、更及时等诸多优势，成为世界各国构建下一代互联网的核心协议。中国则作为推进世界IPv6发展的“引擎”，已经建成目前世界上规模最大的纯IPv6实验骨干网即第二代中国教育和科研计算机网CERNET2，实现了全国100多所重点高校和科研机构的IPv6高速接入，IPv6用户数已超过100万。
高校校园网是高校不可缺少的重要信息化基础设施。可以比喻IPv4是“生产网”，承载着为广大师生提供教学、科研等信息服务；IPv6是“研究网”，担负着为网络新技术试用的科学研究任务。本论文从研究国内外IPv6发展现状出发，就IPv4向IPv6演进过程中所涉及的进程和策略等进行比较深入的分析，系统性地归纳总结并阐明校园网升级为全面支持IPv6所需要解决的核心关键问题、策略和可行的技术路线。并以重庆大学万兆IPv4/IPv6双栈网为例，其IPv6流量已经超过IPv4，核心骨干网采用双链路双归接入技术策略，以实现设备和链路级冗余，确保了网络的稳定性。双栈骨干网路由采用OSPF路由协议，核心层节点之间直接启用OSPF路由协议，分成多个区域以保证路由的收敛速度。IPv6地址规划依据地理位置进行大地址聚合，便于故障位置的追朔和快速定位。升级网络管理、安全监控和认证计费等运行支撑系统，保证升级后的校园网可管、可控和安全。建设一个高速的纯IPv6专网，用于高性能和云计算机、科学研究和IPv6实验研究等。通过设置IPv4/IPv6双栈DNS域名解析系统，利用域名配置的BIND机制，保证用户实现“IPv6访问优先”。
对于IPv4/IPv6双栈网络，如何科学评估2个同时运行的网络的稳定性和可靠性，需要正确测量反映网络运行和性能的延迟、丢包率和带宽等数据。作者收集了重庆大学校园网近2年来IPv4/IPv6双栈网接入、汇聚和骨干网的延迟、丢包率等实际运行数据。同时结合IPv6报文的新特点，利用扩展报头和未定义的扩展字段，提出并设计一个IPv6网络延迟和丢包率的混合测量模型，并研发出原型系统，在IPv4/IPv6网中进行了适当范围内的部署。通过数据对比分析，该模型利用IPv6报文的特点所设计的模型是科学合理的，对于提高IPv6网络延迟、丢包率测量的精确度有较好帮助。
网络带宽是衡量网络使用情况的一个重要指标。论文较为深入地理解和分析网络带宽测量基本原理，对比了带宽测量典型模型及算法和方法。系统分析和研究IPv6协议的新特点，结合带宽测量的基本原理，在包对带宽测量算法的基础上，提出基于IPv6网络的单程不等长包对减法带宽测量算法，并对IPv4网络向IPv6网络演进的过程中，普遍存在的隧道和互联互通机制，作了算法的处理。该算法减少了发送探测分组的数目；降低了数据包排队带来的影响；避免回程引起的路由对称性。在深入剖析IPv6协议的基础上，设计出基于IPv6的算法实现方案，创新性地提出利用IPv6报头的流标签来实现测量报文序列的严格路径控制方法；利用专门的网络模拟软件OPNET来实现算法的仿真和验证，以证明改进后算法的科学和合理性。
After the birth of the Internet based on the IPv4over forty years ago, such primaryproblems as the address shortage and the Internet security are gradually turning up.Therefore, the IPv6is going to serve as the core protocol for the world’s next-generationInternet because it is more extensive, faster, safer, and timelier than the IPv4. As the“engine” to drive the global development of the IPv6, China has constructed the largestsole IPv6experimental network backbone in the world---the second generation ofChina Education and Research Network-CERNET2, which materializes the high-speednext-generation Internet IPv6access of100-plus key colleges&universities andscientific institutions throughout China, and the IPv6users have outnumbered1million.
The campus network within colleges and universities is an indispensableinformation infrastructure. In the campus network, the IPv4works as the “productionnetwork” to provide teachers and students with information of teaching, scientificresearch, etc., while the IPv6is the “research network” to try the new networktechnology for the scientific research. On the basis of the research on the currentdevelopment of the IPv6both nationally and internationally, the course, the strategy, etc.involved in the evolution of the IPv4to the IPv6are to be intensively analyzed, and thecritical problem, the strategy and the feasible technology in the upgrade of the campusnetwork to completely support the IPv6are to be systematically summarized andclarified. Here Chongqing University which uses a10G IPv4/IPv6dual-stack network istaken as an example. The traffic of the IPv6has outstripped that of the IPv4, and thecore network backbone adopting the dual-link and dual-homing access technology routerealizes the redundancy structure of the equipment and the link level and ensures thestability of the network. The dual-stack network employs the OSPF routing protocolcomprehensively, which is directly enabled between the nodes of the core layer, andvarious areas are divided into for the convergence speed of the route. The IPv6networkplans the main address convergence according to the geographic location to facilitatethe fault location and the quick positioning. The running support systems of the networkmanagement, the safety supervision, the authentication charging, etc. are to be promotedto guarantee the manageability, controllability and safety of the upgraded campusnetwork. A high-speed sole IPv6network will be constructed for the high-performancecalculation, cloud technology, scientific research, and IPv6experimental study. The “IPv6access priority” of users is achieved by the establishment of the IPv4/IPv6dual-stack DNS domain name resolution system and the utilization of the BINDmechanism allocated to the domain name.
For the IPv4/IPv6dual-stack network, assessing the stability and reliability of thetwo networks operating simultaneously in a scientific way depends uponcorrectly-measured data like delay, rate of request timed out and bandwidth whichindicate the operation and performance of the networks. In this way, the recent twoyears’ actual operation data of the IPv4/IPv6dual-stack network of ChongqingUniversity are collected, including the access, the convergence, and the delay and rateof request timed out of the network backbone. Meanwhile, the new characteristics of theIPv6message are utilized to create a new usage of its extension header and undefinedextended field. Then a mixing measurement model of the delay and rate of requesttimed out of the IPv6network is brought forward and designed, and a prototype systemis developed so as to make a deployment within a certain scope in the IPv4/v6network.The comparison with the actual data demonstrates that the model designed on theutilized new characteristics of the IPv6message is reasonable and scientific, which canfacilitate improving the accurate measurement of the delay and rate of request timed outof the IPv6network.
The network bandwidth is a key indicator for the usage of the network. By virtueof relatively thorough understanding and analysis of the basic measurement principle ofthe network bandwidth, an analogy between the typical bandwidth measurement model,algorithm and method is carried out. Then an improved algorithm is presented for theone-way bandwidth measurement of the IPv6-based unequal-packet-pair subtraction, onthe basis of the systematically-analyzed and researched IPv6’s new characteristics, thebasic bandwidth measurement principle, as well as the packet pair bandwidthmeasurement algorithm. In addition, as for the ordinary tunnels and the interconnectionmechanism during the long evolution from the IPv4to the IPv6, the algorithm isprocessed. As a result of the improved algorithm, the number of the detection groups isreduced, the impact imposed by data packet queuing is weakened and the routesymmetry caused by the return trip is avoided. Through further understanding of theIPv6, a plan is designed to achieve the IPv6-based algorithm and the strict path controlfor message sequence measuring is also realized by innovative adoption of the flowlabel of the IPv6header. Finally it turns out that the improved algorithm is scientific andreasonable through algorithm stimulation and confirmation on the specialized network simulation software, OPNET.
高校校园网是高校不可缺少的重要信息化基础设施。可以比喻IPv4是“生产网”，承载着为广大师生提供教学、科研等信息服务；IPv6是“研究网”，担负着为网络新技术试用的科学研究任务。本论文从研究国内外IPv6发展现状出发，就IPv4向IPv6演进过程中所涉及的进程和策略等进行比较深入的分析，系统性地归纳总结并阐明校园网升级为全面支持IPv6所需要解决的核心关键问题、策略和可行的技术路线。并以重庆大学万兆IPv4/IPv6双栈网为例，其IPv6流量已经超过IPv4，核心骨干网采用双链路双归接入技术策略，以实现设备和链路级冗余，确保了网络的稳定性。双栈骨干网路由采用OSPF路由协议，核心层节点之间直接启用OSPF路由协议，分成多个区域以保证路由的收敛速度。IPv6地址规划依据地理位置进行大地址聚合，便于故障位置的追朔和快速定位。升级网络管理、安全监控和认证计费等运行支撑系统，保证升级后的校园网可管、可控和安全。建设一个高速的纯IPv6专网，用于高性能和云计算机、科学研究和IPv6实验研究等。通过设置IPv4/IPv6双栈DNS域名解析系统，利用域名配置的BIND机制，保证用户实现“IPv6访问优先”。
对于IPv4/IPv6双栈网络，如何科学评估2个同时运行的网络的稳定性和可靠性，需要正确测量反映网络运行和性能的延迟、丢包率和带宽等数据。作者收集了重庆大学校园网近2年来IPv4/IPv6双栈网接入、汇聚和骨干网的延迟、丢包率等实际运行数据。同时结合IPv6报文的新特点，利用扩展报头和未定义的扩展字段，提出并设计一个IPv6网络延迟和丢包率的混合测量模型，并研发出原型系统，在IPv4/IPv6网中进行了适当范围内的部署。通过数据对比分析，该模型利用IPv6报文的特点所设计的模型是科学合理的，对于提高IPv6网络延迟、丢包率测量的精确度有较好帮助。
网络带宽是衡量网络使用情况的一个重要指标。论文较为深入地理解和分析网络带宽测量基本原理，对比了带宽测量典型模型及算法和方法。系统分析和研究IPv6协议的新特点，结合带宽测量的基本原理，在包对带宽测量算法的基础上，提出基于IPv6网络的单程不等长包对减法带宽测量算法，并对IPv4网络向IPv6网络演进的过程中，普遍存在的隧道和互联互通机制，作了算法的处理。该算法减少了发送探测分组的数目；降低了数据包排队带来的影响；避免回程引起的路由对称性。在深入剖析IPv6协议的基础上，设计出基于IPv6的算法实现方案，创新性地提出利用IPv6报头的流标签来实现测量报文序列的严格路径控制方法；利用专门的网络模拟软件OPNET来实现算法的仿真和验证，以证明改进后算法的科学和合理性。
After the birth of the Internet based on the IPv4over forty years ago, such primaryproblems as the address shortage and the Internet security are gradually turning up.Therefore, the IPv6is going to serve as the core protocol for the world’s next-generationInternet because it is more extensive, faster, safer, and timelier than the IPv4. As the“engine” to drive the global development of the IPv6, China has constructed the largestsole IPv6experimental network backbone in the world---the second generation ofChina Education and Research Network-CERNET2, which materializes the high-speednext-generation Internet IPv6access of100-plus key colleges&universities andscientific institutions throughout China, and the IPv6users have outnumbered1million.
The campus network within colleges and universities is an indispensableinformation infrastructure. In the campus network, the IPv4works as the “productionnetwork” to provide teachers and students with information of teaching, scientificresearch, etc., while the IPv6is the “research network” to try the new networktechnology for the scientific research. On the basis of the research on the currentdevelopment of the IPv6both nationally and internationally, the course, the strategy, etc.involved in the evolution of the IPv4to the IPv6are to be intensively analyzed, and thecritical problem, the strategy and the feasible technology in the upgrade of the campusnetwork to completely support the IPv6are to be systematically summarized andclarified. Here Chongqing University which uses a10G IPv4/IPv6dual-stack network istaken as an example. The traffic of the IPv6has outstripped that of the IPv4, and thecore network backbone adopting the dual-link and dual-homing access technology routerealizes the redundancy structure of the equipment and the link level and ensures thestability of the network. The dual-stack network employs the OSPF routing protocolcomprehensively, which is directly enabled between the nodes of the core layer, andvarious areas are divided into for the convergence speed of the route. The IPv6networkplans the main address convergence according to the geographic location to facilitatethe fault location and the quick positioning. The running support systems of the networkmanagement, the safety supervision, the authentication charging, etc. are to be promotedto guarantee the manageability, controllability and safety of the upgraded campusnetwork. A high-speed sole IPv6network will be constructed for the high-performancecalculation, cloud technology, scientific research, and IPv6experimental study. The “IPv6access priority” of users is achieved by the establishment of the IPv4/IPv6dual-stack DNS domain name resolution system and the utilization of the BINDmechanism allocated to the domain name.
For the IPv4/IPv6dual-stack network, assessing the stability and reliability of thetwo networks operating simultaneously in a scientific way depends uponcorrectly-measured data like delay, rate of request timed out and bandwidth whichindicate the operation and performance of the networks. In this way, the recent twoyears’ actual operation data of the IPv4/IPv6dual-stack network of ChongqingUniversity are collected, including the access, the convergence, and the delay and rateof request timed out of the network backbone. Meanwhile, the new characteristics of theIPv6message are utilized to create a new usage of its extension header and undefinedextended field. Then a mixing measurement model of the delay and rate of requesttimed out of the IPv6network is brought forward and designed, and a prototype systemis developed so as to make a deployment within a certain scope in the IPv4/v6network.The comparison with the actual data demonstrates that the model designed on theutilized new characteristics of the IPv6message is reasonable and scientific, which canfacilitate improving the accurate measurement of the delay and rate of request timed outof the IPv6network.
The network bandwidth is a key indicator for the usage of the network. By virtueof relatively thorough understanding and analysis of the basic measurement principle ofthe network bandwidth, an analogy between the typical bandwidth measurement model,algorithm and method is carried out. Then an improved algorithm is presented for theone-way bandwidth measurement of the IPv6-based unequal-packet-pair subtraction, onthe basis of the systematically-analyzed and researched IPv6’s new characteristics, thebasic bandwidth measurement principle, as well as the packet pair bandwidthmeasurement algorithm. In addition, as for the ordinary tunnels and the interconnectionmechanism during the long evolution from the IPv4to the IPv6, the algorithm isprocessed. As a result of the improved algorithm, the number of the detection groups isreduced, the impact imposed by data packet queuing is weakened and the routesymmetry caused by the return trip is avoided. Through further understanding of theIPv6, a plan is designed to achieve the IPv6-based algorithm and the strict path controlfor message sequence measuring is also realized by innovative adoption of the flowlabel of the IPv6header. Finally it turns out that the improved algorithm is scientific andreasonable through algorithm stimulation and confirmation on the specialized network simulation software, OPNET.
引文
[1]Baran P. On Distibuteed Commnications Networks[J].IEEE Transactions on CommunicationSystems,1964,Vol.12(1):1-9.
[2]Lukasik,et al..Why the Arpanet Was Built[J], IEEE Annals of The History of Computing,2011,Vol.33(3):4-20.
[3]Al Zahr, Sawsan,et al..Physical Layer Impairments in WDM Core Networks: a Comparisonbetween a North-American Backbone and a Pan-European Backbone[J],2ndInternationalConference on Broadband Network,2005,335-339.
[4]Bagnulo Marcelo,et al.. How to Contribute Research Results to Internet Standardization[J],Computer Communication Review,2011,Vol41(3):10-13.
[5]Kantarci Burak,et al..Performance Analysis of SLA-Aware Connection Provisioning in OpticalNetworks [J],13thInternational Conference on Transparent Optical Network(ICTON),2011.
[6]Wu Jianping,et al..CNGI－CERNET2: an IPv6Deployment in China [J],ComputerCommunication Review,2011,Vol41(2):48-52.
[7]Yang Jia-Hai,et al..Towards Next Generation Internet Management:CNGICERNET2Experiences[J],Journal of Computer Science and Technology,2009, Vol24(3):482-494.
[8]张光妲，基于Ipv6的下一代校园网的过渡研究[D].哈尔滨工业大学.哈尔滨.2007.
[9]Morelli M,et al.. An IPv6Internet Exchange model. Lessons from Euro6IX project[J],2005Symosium on Applictions and the Internat Workshops,Proceedings,2005,50-53.
[10]Chown T,at al..Results and plans of the IPv6test-bed initiatives within the Europeancommission IST programme[J],2003Symosium on Applictions and the InternatWorkshops,Proceedings,2003,145-148.
[11]工业与信息化部电信电信研究院,跟踪研究,日本IPv6的发展进程[J],2009,vol(24),http://wenku.baidu.com/view/b5232f3543323968011c92df.html
[12]6Bone-KR,http://www.6bone.ne.kr.
[13]Wang JiLong,at al.. DRAGON-Lab-Next generation internet technology experiment platform [J],Science in China Series F-information Sciences,2008,Vol.51(11):1908-1918.
[14]徐小兴.运营商IPv6演进实践与思考[EB/01].现代电信科技,http://www.c114.net/topic/2772/a594119.html.2011.7.
[15]刘春林.下一代互联网演进技术分析[EB/01],电信快报，http://www.cnki.com.cn/Article/CJFDTOTAL-DXKB201107004.htm,2011.7.
[16]史凡.等，下一代互联网体系架构与现网演进实践[EB/01]，电信技术，http://www.ttm.com.cn/article/article.asp?id=142291&bid=8296,2011.4
[17]张伟平.等， IPv6业务网络演进方案分析[EB/01]，电信技术，http://www.ttm.com.cn/article/article.asp?id=142293&bid=8296,2011.4.
[18]杨家海.等,互联网络测量理论与应用[M],人民邮电出版社,2009.
[19]A.McGregor,at al..the NLANR network analysis infrastructure[J].IEEE CommunicationsMahazine.2000,38(5):122-128
[20]Paxson.V,Almes.G,et al.An Architecture for large-scale Internet Measurement[J].IEEEComuications,1998,36(8):48-54.
[21] IETF[EB/OL]. http://www.ietf.org
[22]RIPE NCC. http://www.ripe.net/ripe/docs/ripe-214.html.2000
[23]SURVEYOR[EB/OL].http://www.advanced.org/survey/
[24]孙红杰，基于主动测量的网络性能分析[D].哈尔滨工业大学.哈尔滨.2007.
[25]曾彬，基于主动测试的网络性能监测技术研究[D].湖南大学.长沙.2009.
[26]王继龙,等.大规模计算机网络性能监控模型的设计与实现[J].计算机研究与发展,2000,Vol.37(4):443-452.
[27]毕经平,等. Internet端到端性能监测与分析系统的关键技术研究[J].计算机辅助设计与图形学学报,2003, Vol.15(01):111-117.
[28]Gaogang XIE, Jianhua YANG, Valerie ISSARNY, et al. An Accurate and Efficient3-PhaseMeasurement Method for IP Traffic Flow on High Speed Link[J].In: Proc of the2007IEEEInternational Conference on Networking, Martinique, French Caribbean,2007,46-46.
[29]黎文伟,等.基于通用PC架构的高精度网络时延测量方法[J].软件学报,2006, Vol.17(02):275-284.
[30]张文杰,等.互联网应用性能测量系统的研究与实现[J].计算机研究与发展,2003, Vol.40(01):60-67
[31]张文杰,等.bing算法估测网络带宽的研究与实现[J],小型微型计算机系统.2004.vol.25(5).833-835.
[32]张文杰.可定制的网络测量基础设施研究[D].西安交通大学.西安.2002.
[33]张宏莉,等.Internet测量与分析综述[J],软件学报,2003.Vol14(1).110-116.
[34]姜誉,等.大型ISP网络拓扑多点测量及其特征分析实例[J].软件学报,2005,Vol.16(05):846-856.
[35]蔡志平,等.网络测量部署模型及其优化算法[J].软件学报,2008,Vol.19(02):419-431
[36]王俊峰,等.网络测量中自适应数据采集方法[J].软件学报,2004,Vol.15(08):1227-1236.
[37]林宇,等.IP网端到端性能测量技术研究的进展[J].电子学报.2003,Vol.31(8).1227-1233.
[38]林宇,等.一种非对称链路带宽测量算法[J].电子学报.2002.vol.30(12A).2065-2070.
[39]王立超,IPv4/v6过渡技术的研究与应用[D].重庆大学.重庆.2010.
[40] Douglas E.Comer.用TCP/IP进行网际互连[M],北京:电子工业出版社.2001.
[41]Naegele.Jackson S,at al..IPPM measurements of the German research network G-WiN and theirapplication to videoconferencing services[J],Proceedings of the Eighth IASTED InternationalConference on Internet and Multimedia Systems and Applications,2004,317-322.
[42]Naegele Jackson.S,at al..IPPM measurements and network load behavior of the GermanResearch Network G-WiN[J],International Conference on Computing, Communications and ControlTechnologies,2004, Vol.3,390-395.
[43]赵洪华,等.网络性能特性的描述和测量[J].解放军理工大学学报.2004.Vol.5(5).19-21.
[44]Vern Paxson.Measurements and Analysis of End-to-End Internet Dynamics[D],Berkeley,1997:12-20.
[45]Vern Paxson,et al..An Architecture for Large-Scale Internet Measurement[J].IEEECommunications,1998,36(8):48-54.
[46]陈松,互联网络测量管理若干关键技术研究[D].电子科技大学.成都.2010.
[47]Attila Pasztor,at al.. A precision infrastructure for active probing. In:Proc of Passive and ActiveMeasurement Workshop (PAM2001), Amsterdam,Netherlands, April2001,33-44.
[48] Cisco. http://www.cisco.com/go/netflow.
[49]Attila Pasztor,at al.. On the Scope of End-to-End Probing Methods.IEEE CommunicationsLetters,2002,6(11):509–511.
[50]Les Cottrell.Comparison of some Internet Active End-to-end Performance Measurement projects.http://www.slac.stanford.edu/comp/net/wan-mon/iepm-cf.html,1999-7-5
[51] CAIDA, Tools of CAIDA. http://www.caida.org/tools/,2008-12-22
[52]J Case, et al.. Management Information Base for Version2of the Simple Network ManagementProtocol (SNMPv2), RFC1907, Internet Engineering Task Force,1996
[53] H Burch, at al.. Mapping the Internet[J]. IEEE Computer,1999,32(4):97-98
[54]金勇进,蒋妍,李序颖.抽样技术[M]，北京中国人民大学出版社,2002
[55]胡健颖,孙山泽.抽样调查的理论、方法和应用[M]，北京:北京大学出版社,2000..
[56] Mahdavi J,at al.. IPPM Metrics for Measuring Connectivity. RFC2678, InternetEngineering Task Force, Sept,1999.
[57]Kevin L,at al.. Measuring Link Bandwidths Using a Deterministic Model of Packet Delay[J].Proceedings of ACM SIGCOMM2000. Stockholm: ACM Press,2000,283-294.
[58]Prasad R S, et al.. Bandwidth Estimation: Metrics, Measurement,Techniques, and Tools[J]. IEEENetworks,2003,17(6):27-35.
[59] Almes G,at al.. A One-way Round-trip Metric for IPPM. RFC2681,Internet Engineering Task Force, Sept,1999.
[60]王恺, Internet包延迟行为分析和主动测量技术研究[D],中国科学院研究生院（计算技术研究所）,北京,2005.
[61]黄廷辉,IPv6在嵌入式实时系统中的设计与实现[D],吉林大学硕士论文.长春.2006.
[62]王相林著.IPv6核心技术[M],北京:科学出版社,2009.
[63] Silvia Hagen著.技桥译.IPv6精髓[M],北京:清华大学出版社,2004.
[64]Pei Changxing,at al..AB-cap: A Fast Approach to Available Bandwidth Estimation[J],Korea-China Joint Conference on Communication&Information Technology,2004, Mar.17-19, Kyungpook Univ., Daegu, Korea,25-28.
[65]Paxson.Vein,at al..Wide-area traffic:the failure of poisson modeling[J].IEEE/ACM Transactionson Networking,1995,Vol.3(3):226-244.
[66]邱全杰,基于IPv6的网络带宽测量研究[D],重庆大学硕士论文.重庆.2005.
[67]Jain.M,at al..End-to-end available bandwidth:Measurement methodology,dynamics,and relationwith TCP throughout[C],Proceedings of ACM S1GCOMM,2002.
[68]E. Goldoni,at al..End-to-End Available Band-width Estimation Tools, An ExperimentalComparison[J],In Proceedings of Traffic Monitoring and Analysis,2010,71-82.
[69]邱全杰,等.一种IPv6网络可用带宽测量方法及分析[J],计算机科学,2011,Vol.38(3)84-86,121.
[70] Quan-jie Qiu,at al..The End-to-end IPv6Network Available Bandwidth Measurement andAnalysis[J],International Journal of Advancements in Computing Technology,2012,Vol.4(5)256~264.
[71]Quanjie Qiu,at al..algorithm for available bandwidth measurement of IPv6network[J], TheInternational Conference on Advanced Data Mining and Applications,2010,419-426.
[72]R.L. Carter, M.E. Crovella.Server selection using dynamic path characterization in wide-areanetworks[J].In Proceedings IEEE INFOCOM '97,1997,vol(3):1014-1021.
[73] K. Lai, at al..Measuring bandwidth[J]. In Proceedings IEEE INFOCOM '99,1999,235-245
[74]C.Dovrolis, P.at al..What do packet dispersion techniques measure[J], In Proceedings IEEEINFOCOM,2001, vol.(2):905-914.
[75]Nitul Dutta,at al..Mathematical Analysis of Signaling Overhead in MIPv6Based N-LayerArchitecture[J], JCIT: Journal of Convergence Information Technology,2010,vol.5(8):pp.252-261.
[76]Cesar D. Guerrero,at al..On the applicability of available bandwidth estimation techniques andtools[J],Computer Communications,2010,vol.33(1):11-22.
[77]Yu Lin, Haitao Wu.Using a Multi-packet model to Measure Asymmetric Link Bandwidth inInternet[J].IEEE ICC`03,Anchorage, USA, May2003.
[78]R. Prasad,at al..Bandwidth estimation: metrics, measurement techniques and tools[J], IEEENetwork,2003,vol.17(6):27-35.
[79]刘星成,等.网络可用带宽的高精度测量算法[J].电子学报,2007,vol.35(1):68-72.
[80]Alouf.S,at al..Inferring network characteristics via moment-based estimators[C].Anichor-ageAlaska,USA:The20th Annual Joint Conf on the IEEE Computer and CommunicationsSocieties,2001.
[81]Hei,XJ.at al.. A light-weight available band-width inference methodology in a queueing analysisapproach[J].Proe of the40th IEEE Int'l Conf on Communications.Los Alamitos,CA:IEEE ComputerSociety Press,2005:120-124.
[82]Chen Yu-shi,at al..The improvement of measuring in available bandwidth based onpathload[J],Journal of Xiamen University (Natural Science),2011,Vol.50(1):38-41.
[83]Ribeiro.V, et al..PathChirp:Etticient available bandwidth egimmion for network paths[J].SanDiego,CA:Proc of Passive and Active Measurements Workshop,2003.
[84]Jain.M,at al.. A measurement tool for end-to-end available bandwidth[J].FortCollins,CO,USA:Proc of Passive and Active Measurements Workshop,2002,14-25.
[85]陈敏.Opnet网络仿真[M],北京:清华大学出版社,2004.
[2]Lukasik,et al..Why the Arpanet Was Built[J], IEEE Annals of The History of Computing,2011,Vol.33(3):4-20.
[3]Al Zahr, Sawsan,et al..Physical Layer Impairments in WDM Core Networks: a Comparisonbetween a North-American Backbone and a Pan-European Backbone[J],2ndInternationalConference on Broadband Network,2005,335-339.
[4]Bagnulo Marcelo,et al.. How to Contribute Research Results to Internet Standardization[J],Computer Communication Review,2011,Vol41(3):10-13.
[5]Kantarci Burak,et al..Performance Analysis of SLA-Aware Connection Provisioning in OpticalNetworks [J],13thInternational Conference on Transparent Optical Network(ICTON),2011.
[6]Wu Jianping,et al..CNGI－CERNET2: an IPv6Deployment in China [J],ComputerCommunication Review,2011,Vol41(2):48-52.
[7]Yang Jia-Hai,et al..Towards Next Generation Internet Management:CNGICERNET2Experiences[J],Journal of Computer Science and Technology,2009, Vol24(3):482-494.
[8]张光妲，基于Ipv6的下一代校园网的过渡研究[D].哈尔滨工业大学.哈尔滨.2007.
[9]Morelli M,et al.. An IPv6Internet Exchange model. Lessons from Euro6IX project[J],2005Symosium on Applictions and the Internat Workshops,Proceedings,2005,50-53.
[10]Chown T,at al..Results and plans of the IPv6test-bed initiatives within the Europeancommission IST programme[J],2003Symosium on Applictions and the InternatWorkshops,Proceedings,2003,145-148.
[11]工业与信息化部电信电信研究院,跟踪研究,日本IPv6的发展进程[J],2009,vol(24),http://wenku.baidu.com/view/b5232f3543323968011c92df.html
[12]6Bone-KR,http://www.6bone.ne.kr.
[13]Wang JiLong,at al.. DRAGON-Lab-Next generation internet technology experiment platform [J],Science in China Series F-information Sciences,2008,Vol.51(11):1908-1918.
[14]徐小兴.运营商IPv6演进实践与思考[EB/01].现代电信科技,http://www.c114.net/topic/2772/a594119.html.2011.7.
[15]刘春林.下一代互联网演进技术分析[EB/01],电信快报，http://www.cnki.com.cn/Article/CJFDTOTAL-DXKB201107004.htm,2011.7.
[16]史凡.等，下一代互联网体系架构与现网演进实践[EB/01]，电信技术，http://www.ttm.com.cn/article/article.asp?id=142291&bid=8296,2011.4
[17]张伟平.等， IPv6业务网络演进方案分析[EB/01]，电信技术，http://www.ttm.com.cn/article/article.asp?id=142293&bid=8296,2011.4.
[18]杨家海.等,互联网络测量理论与应用[M],人民邮电出版社,2009.
[19]A.McGregor,at al..the NLANR network analysis infrastructure[J].IEEE CommunicationsMahazine.2000,38(5):122-128
[20]Paxson.V,Almes.G,et al.An Architecture for large-scale Internet Measurement[J].IEEEComuications,1998,36(8):48-54.
[21] IETF[EB/OL]. http://www.ietf.org
[22]RIPE NCC. http://www.ripe.net/ripe/docs/ripe-214.html.2000
[23]SURVEYOR[EB/OL].http://www.advanced.org/survey/
[24]孙红杰，基于主动测量的网络性能分析[D].哈尔滨工业大学.哈尔滨.2007.
[25]曾彬，基于主动测试的网络性能监测技术研究[D].湖南大学.长沙.2009.
[26]王继龙,等.大规模计算机网络性能监控模型的设计与实现[J].计算机研究与发展,2000,Vol.37(4):443-452.
[27]毕经平,等. Internet端到端性能监测与分析系统的关键技术研究[J].计算机辅助设计与图形学学报,2003, Vol.15(01):111-117.
[28]Gaogang XIE, Jianhua YANG, Valerie ISSARNY, et al. An Accurate and Efficient3-PhaseMeasurement Method for IP Traffic Flow on High Speed Link[J].In: Proc of the2007IEEEInternational Conference on Networking, Martinique, French Caribbean,2007,46-46.
[29]黎文伟,等.基于通用PC架构的高精度网络时延测量方法[J].软件学报,2006, Vol.17(02):275-284.
[30]张文杰,等.互联网应用性能测量系统的研究与实现[J].计算机研究与发展,2003, Vol.40(01):60-67
[31]张文杰,等.bing算法估测网络带宽的研究与实现[J],小型微型计算机系统.2004.vol.25(5).833-835.
[32]张文杰.可定制的网络测量基础设施研究[D].西安交通大学.西安.2002.
[33]张宏莉,等.Internet测量与分析综述[J],软件学报,2003.Vol14(1).110-116.
[34]姜誉,等.大型ISP网络拓扑多点测量及其特征分析实例[J].软件学报,2005,Vol.16(05):846-856.
[35]蔡志平,等.网络测量部署模型及其优化算法[J].软件学报,2008,Vol.19(02):419-431
[36]王俊峰,等.网络测量中自适应数据采集方法[J].软件学报,2004,Vol.15(08):1227-1236.
[37]林宇,等.IP网端到端性能测量技术研究的进展[J].电子学报.2003,Vol.31(8).1227-1233.
[38]林宇,等.一种非对称链路带宽测量算法[J].电子学报.2002.vol.30(12A).2065-2070.
[39]王立超,IPv4/v6过渡技术的研究与应用[D].重庆大学.重庆.2010.
[40] Douglas E.Comer.用TCP/IP进行网际互连[M],北京:电子工业出版社.2001.
[41]Naegele.Jackson S,at al..IPPM measurements of the German research network G-WiN and theirapplication to videoconferencing services[J],Proceedings of the Eighth IASTED InternationalConference on Internet and Multimedia Systems and Applications,2004,317-322.
[42]Naegele Jackson.S,at al..IPPM measurements and network load behavior of the GermanResearch Network G-WiN[J],International Conference on Computing, Communications and ControlTechnologies,2004, Vol.3,390-395.
[43]赵洪华,等.网络性能特性的描述和测量[J].解放军理工大学学报.2004.Vol.5(5).19-21.
[44]Vern Paxson.Measurements and Analysis of End-to-End Internet Dynamics[D],Berkeley,1997:12-20.
[45]Vern Paxson,et al..An Architecture for Large-Scale Internet Measurement[J].IEEECommunications,1998,36(8):48-54.
[46]陈松,互联网络测量管理若干关键技术研究[D].电子科技大学.成都.2010.
[47]Attila Pasztor,at al.. A precision infrastructure for active probing. In:Proc of Passive and ActiveMeasurement Workshop (PAM2001), Amsterdam,Netherlands, April2001,33-44.
[48] Cisco. http://www.cisco.com/go/netflow.
[49]Attila Pasztor,at al.. On the Scope of End-to-End Probing Methods.IEEE CommunicationsLetters,2002,6(11):509–511.
[50]Les Cottrell.Comparison of some Internet Active End-to-end Performance Measurement projects.http://www.slac.stanford.edu/comp/net/wan-mon/iepm-cf.html,1999-7-5
[51] CAIDA, Tools of CAIDA. http://www.caida.org/tools/,2008-12-22
[52]J Case, et al.. Management Information Base for Version2of the Simple Network ManagementProtocol (SNMPv2), RFC1907, Internet Engineering Task Force,1996
[53] H Burch, at al.. Mapping the Internet[J]. IEEE Computer,1999,32(4):97-98
[54]金勇进,蒋妍,李序颖.抽样技术[M]，北京中国人民大学出版社,2002
[55]胡健颖,孙山泽.抽样调查的理论、方法和应用[M]，北京:北京大学出版社,2000..
[56] Mahdavi J,at al.. IPPM Metrics for Measuring Connectivity. RFC2678, InternetEngineering Task Force, Sept,1999.
[57]Kevin L,at al.. Measuring Link Bandwidths Using a Deterministic Model of Packet Delay[J].Proceedings of ACM SIGCOMM2000. Stockholm: ACM Press,2000,283-294.
[58]Prasad R S, et al.. Bandwidth Estimation: Metrics, Measurement,Techniques, and Tools[J]. IEEENetworks,2003,17(6):27-35.
[59] Almes G,at al.. A One-way Round-trip Metric for IPPM. RFC2681,Internet Engineering Task Force, Sept,1999.
[60]王恺, Internet包延迟行为分析和主动测量技术研究[D],中国科学院研究生院（计算技术研究所）,北京,2005.
[61]黄廷辉,IPv6在嵌入式实时系统中的设计与实现[D],吉林大学硕士论文.长春.2006.
[62]王相林著.IPv6核心技术[M],北京:科学出版社,2009.
[63] Silvia Hagen著.技桥译.IPv6精髓[M],北京:清华大学出版社,2004.
[64]Pei Changxing,at al..AB-cap: A Fast Approach to Available Bandwidth Estimation[J],Korea-China Joint Conference on Communication&Information Technology,2004, Mar.17-19, Kyungpook Univ., Daegu, Korea,25-28.
[65]Paxson.Vein,at al..Wide-area traffic:the failure of poisson modeling[J].IEEE/ACM Transactionson Networking,1995,Vol.3(3):226-244.
[66]邱全杰,基于IPv6的网络带宽测量研究[D],重庆大学硕士论文.重庆.2005.
[67]Jain.M,at al..End-to-end available bandwidth:Measurement methodology,dynamics,and relationwith TCP throughout[C],Proceedings of ACM S1GCOMM,2002.
[68]E. Goldoni,at al..End-to-End Available Band-width Estimation Tools, An ExperimentalComparison[J],In Proceedings of Traffic Monitoring and Analysis,2010,71-82.
[69]邱全杰,等.一种IPv6网络可用带宽测量方法及分析[J],计算机科学,2011,Vol.38(3)84-86,121.
[70] Quan-jie Qiu,at al..The End-to-end IPv6Network Available Bandwidth Measurement andAnalysis[J],International Journal of Advancements in Computing Technology,2012,Vol.4(5)256~264.
[71]Quanjie Qiu,at al..algorithm for available bandwidth measurement of IPv6network[J], TheInternational Conference on Advanced Data Mining and Applications,2010,419-426.
[72]R.L. Carter, M.E. Crovella.Server selection using dynamic path characterization in wide-areanetworks[J].In Proceedings IEEE INFOCOM '97,1997,vol(3):1014-1021.
[73] K. Lai, at al..Measuring bandwidth[J]. In Proceedings IEEE INFOCOM '99,1999,235-245
[74]C.Dovrolis, P.at al..What do packet dispersion techniques measure[J], In Proceedings IEEEINFOCOM,2001, vol.(2):905-914.
[75]Nitul Dutta,at al..Mathematical Analysis of Signaling Overhead in MIPv6Based N-LayerArchitecture[J], JCIT: Journal of Convergence Information Technology,2010,vol.5(8):pp.252-261.
[76]Cesar D. Guerrero,at al..On the applicability of available bandwidth estimation techniques andtools[J],Computer Communications,2010,vol.33(1):11-22.
[77]Yu Lin, Haitao Wu.Using a Multi-packet model to Measure Asymmetric Link Bandwidth inInternet[J].IEEE ICC`03,Anchorage, USA, May2003.
[78]R. Prasad,at al..Bandwidth estimation: metrics, measurement techniques and tools[J], IEEENetwork,2003,vol.17(6):27-35.
[79]刘星成,等.网络可用带宽的高精度测量算法[J].电子学报,2007,vol.35(1):68-72.
[80]Alouf.S,at al..Inferring network characteristics via moment-based estimators[C].Anichor-ageAlaska,USA:The20th Annual Joint Conf on the IEEE Computer and CommunicationsSocieties,2001.
[81]Hei,XJ.at al.. A light-weight available band-width inference methodology in a queueing analysisapproach[J].Proe of the40th IEEE Int'l Conf on Communications.Los Alamitos,CA:IEEE ComputerSociety Press,2005:120-124.
[82]Chen Yu-shi,at al..The improvement of measuring in available bandwidth based onpathload[J],Journal of Xiamen University (Natural Science),2011,Vol.50(1):38-41.
[83]Ribeiro.V, et al..PathChirp:Etticient available bandwidth egimmion for network paths[J].SanDiego,CA:Proc of Passive and Active Measurements Workshop,2003.
[84]Jain.M,at al.. A measurement tool for end-to-end available bandwidth[J].FortCollins,CO,USA:Proc of Passive and Active Measurements Workshop,2002,14-25.
[85]陈敏.Opnet网络仿真[M],北京:清华大学出版社,2004.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |