基于客户端的网络服务性能测试模型研究与系统支持
|
摘要
当前,网络服务提供商通过Web站点向各类用户提供了大量种类齐全、功能各异的网络服务;为保证广大用户得到完美的服务体验,同时也为了保障服务提供商的商业利益,这些网络服务必须具备很强的服务处理能力和很大的服务容量,能够支持大量用户的并发服务请求;要确定网络服务的性能是否达到了提供商的预期目标,必须运用基于客户端的网络服务性能自动测试工具对其进行全面有效的性能测试,准确地评价出服务性能的优劣,揭示并定位服务中可能存在的性能瓶颈,从而有针对性地进行性能优化,进而达到提高服务性能的目的。
在进行基于客户端的网络服务性能测试过程中,良好的测试模型是整个测试过程能否达到预期目标的前提和基础;测试模型对整个性能测试过程中起着方向性和指导性的作用,它是对整个测试体系的刻画,是对测试目标系统的描述,是对整个应测内容的编排,故测试模型的选用在很大程度上决定了测试过程的有效性;而良好的测试模型来源于正确的模型建立策略。
本文详细分析和研究了现阶段已有的基于客户端的各类性能测试模型所具有的优点和存在的不足,并从中提取好的测试建模思想加以利用;同时细致分析了当前在建立测试模型过程中所使用的相关建模技术。
本文在详尽分析了服务体系中三个组成部分(客户端、网络部分、服务端)对整个服务性能测试模型建立过程所构成的各种影响因素和模型建立过程中所面临的主要问题的基础上,根据各部分所具有的不同特性以及在体系中所起的作用和所占有的地位,将它们区别对待,对不同层次的问题使用不同的抽象方法,提出一种新的服务性能测试模型建立策略——层次化的测试模型建立策略,它是一种可变粒度的事务级建模策略;该策略将测试模型的建立分为两部分,一部分是建立用户访问服务模型,一部分是建立访问流量模型;用户访问服务模型模拟真实用户使用网络服务的方式,而访问流量模型则模拟真实的Web数据流量,使依据该模型所产生的Web负载流量符合真实Web数据流量的各种统计特性,使其具有骤变的、重尾的特性以及相互关联的到达模式;同时详细阐述了该建模策略包含的这两部分模型的具体建立步骤,充分体现了层次化建模的特点,即:层次明了,过程清晰;该建模策略在整体上具有良好的扩展性和灵活性。
依据层次化测试模型建立策略的核心思想,使用Java语言实现了一个性能自动测试工具原型系统的测试建模部分;该测试建模部分通过其与测试执行部分间的接口最终以XML文档的形式为测试执行部分提供完整而详细的目标系统的测试模型信息。
Many different kind of Website Services are provided to network users by Service provider through the users accessing Website, the Website Services afford many different functions to the network users. For guaranteeing the network users getting perfect experience by using the Website Services and at the same time guaranteeing the Services provider getting good profit, these Website Services must have perfect high performance to supporting a lot of network users accessing them subsequently. If the Services providers want to determine if the performances of provided Website Services reach anticipating level and if these Website Services can give good service process to the network users, they must use Client-End automated performance test tool to full-scale testing performances of the Website Services, in order to accurate evaluating the level of performance for these Website Services. At the same time, the process of testing can determine if having bottlenecks of performance in these Website Services and locate where the bottlenecks in these Website Services. Services provider can improve the performance of these Services by solving the bottlenecks of performance in them.During the process of Client-End performance testing for these Website Services, precise model for testing is foundation and precondition of reaching the anticipated testing intention; Model for testing act as direction of the whole process of testing, so it is very important; It is architecture of the whole testing process, is description of target system of testing, is layout of the whole testing content; So choosing what kind of model for testing will make if the process of testing reaching test goal; The precise model for testing comes from right test modeling strategy.This paper analyses and researches on strongpoint and limitation of already existing test models, at the same time get excellent thought of modeling from them and utilize the thought; the paper also all-scale analyses current the techniques of test modeling.This paper particular analyses all kinds of factors impacting the whole modeling process of performance testing for these Services coming from three parts of Website Service architecture(Client-end,Network-part,Service-end) and main problems confronted during the modeling process of performance test; On foundation of the detailed analyse,also basing the three parts having different characteristics and function in the whole Service architecture, using different methods for abstracting for different parts, a new kind of modeling strategy for performance testing of Website Service-layered modeling strategy is proposed; this modeling strategy divides the model of testing into tow parts, one is user accessing Services model, the other is network traffic model; user accessing Services model is used simulates manner of real network users using Website Services, network traffic model simulates real Web traffic in network, enables all kinds of statistical characteristics generated using this model according with all kinds of statistical characteristics real Web traffic in network, makes the generated traffic has changing suddenly and heavy-tailed statistical characteristics.The paper detailed describes approaches of establishing the two parts model in the layered modeling strategy, it sufficient represents the excellent characteristics of this modeling strategy, the excellent characteristics of the modeling strategy are the very perspicuous layer and clear
builded process, the whole of this modeling strategy has high level of flexibility and expansibility.According to the main thought of the layered modeling strategy, using Java programming language to implement a prototype system of a modeling part system for a new Client-End automated performance testing tool; The modeling part system gives all detailed information of performance testing model for testing target system to the performing part system of the Client-End automated performance testing tool by interface between the modeling part and the performing part, the all detailed information contained in performance testing model is saved in one XML document, is expressed as XML format.
在进行基于客户端的网络服务性能测试过程中,良好的测试模型是整个测试过程能否达到预期目标的前提和基础;测试模型对整个性能测试过程中起着方向性和指导性的作用,它是对整个测试体系的刻画,是对测试目标系统的描述,是对整个应测内容的编排,故测试模型的选用在很大程度上决定了测试过程的有效性;而良好的测试模型来源于正确的模型建立策略。
本文详细分析和研究了现阶段已有的基于客户端的各类性能测试模型所具有的优点和存在的不足,并从中提取好的测试建模思想加以利用;同时细致分析了当前在建立测试模型过程中所使用的相关建模技术。
本文在详尽分析了服务体系中三个组成部分(客户端、网络部分、服务端)对整个服务性能测试模型建立过程所构成的各种影响因素和模型建立过程中所面临的主要问题的基础上,根据各部分所具有的不同特性以及在体系中所起的作用和所占有的地位,将它们区别对待,对不同层次的问题使用不同的抽象方法,提出一种新的服务性能测试模型建立策略——层次化的测试模型建立策略,它是一种可变粒度的事务级建模策略;该策略将测试模型的建立分为两部分,一部分是建立用户访问服务模型,一部分是建立访问流量模型;用户访问服务模型模拟真实用户使用网络服务的方式,而访问流量模型则模拟真实的Web数据流量,使依据该模型所产生的Web负载流量符合真实Web数据流量的各种统计特性,使其具有骤变的、重尾的特性以及相互关联的到达模式;同时详细阐述了该建模策略包含的这两部分模型的具体建立步骤,充分体现了层次化建模的特点,即:层次明了,过程清晰;该建模策略在整体上具有良好的扩展性和灵活性。
依据层次化测试模型建立策略的核心思想,使用Java语言实现了一个性能自动测试工具原型系统的测试建模部分;该测试建模部分通过其与测试执行部分间的接口最终以XML文档的形式为测试执行部分提供完整而详细的目标系统的测试模型信息。
Many different kind of Website Services are provided to network users by Service provider through the users accessing Website, the Website Services afford many different functions to the network users. For guaranteeing the network users getting perfect experience by using the Website Services and at the same time guaranteeing the Services provider getting good profit, these Website Services must have perfect high performance to supporting a lot of network users accessing them subsequently. If the Services providers want to determine if the performances of provided Website Services reach anticipating level and if these Website Services can give good service process to the network users, they must use Client-End automated performance test tool to full-scale testing performances of the Website Services, in order to accurate evaluating the level of performance for these Website Services. At the same time, the process of testing can determine if having bottlenecks of performance in these Website Services and locate where the bottlenecks in these Website Services. Services provider can improve the performance of these Services by solving the bottlenecks of performance in them.During the process of Client-End performance testing for these Website Services, precise model for testing is foundation and precondition of reaching the anticipated testing intention; Model for testing act as direction of the whole process of testing, so it is very important; It is architecture of the whole testing process, is description of target system of testing, is layout of the whole testing content; So choosing what kind of model for testing will make if the process of testing reaching test goal; The precise model for testing comes from right test modeling strategy.This paper analyses and researches on strongpoint and limitation of already existing test models, at the same time get excellent thought of modeling from them and utilize the thought; the paper also all-scale analyses current the techniques of test modeling.This paper particular analyses all kinds of factors impacting the whole modeling process of performance testing for these Services coming from three parts of Website Service architecture(Client-end,Network-part,Service-end) and main problems confronted during the modeling process of performance test; On foundation of the detailed analyse,also basing the three parts having different characteristics and function in the whole Service architecture, using different methods for abstracting for different parts, a new kind of modeling strategy for performance testing of Website Service-layered modeling strategy is proposed; this modeling strategy divides the model of testing into tow parts, one is user accessing Services model, the other is network traffic model; user accessing Services model is used simulates manner of real network users using Website Services, network traffic model simulates real Web traffic in network, enables all kinds of statistical characteristics generated using this model according with all kinds of statistical characteristics real Web traffic in network, makes the generated traffic has changing suddenly and heavy-tailed statistical characteristics.The paper detailed describes approaches of establishing the two parts model in the layered modeling strategy, it sufficient represents the excellent characteristics of this modeling strategy, the excellent characteristics of the modeling strategy are the very perspicuous layer and clear
builded process, the whole of this modeling strategy has high level of flexibility and expansibility.According to the main thought of the layered modeling strategy, using Java programming language to implement a prototype system of a modeling part system for a new Client-End automated performance testing tool; The modeling part system gives all detailed information of performance testing model for testing target system to the performing part system of the Client-End automated performance testing tool by interface between the modeling part and the performing part, the all detailed information contained in performance testing model is saved in one XML document, is expressed as XML format.
引文
[1] AOL. America Online Press Data Points.2002. http://corp.aol.com/press/press_datapoints.html.
[2] Zona Research Inc. The economic impacts of unacceptable web-site downloadspeed. 1999.http://www.keynote.com/downloads/whitepapers/economic_impact_of_downlaodspeed.pdf.
[3] ARudolf, RPirker. E-Business testing: User perceptions and performance issues.In: Proc of the 1st Asia Pacific Conf on Quality Software. LosAlamitos: IEEE Computer Society Press, 2000 315～323.
[4] SPECweb99 Release Documents 1.02 July 21,.2000. http://open.spec.org/web99/docs/whitepaper.html
[5] 查志琴,李慧.HTTP协议的发展对Web服务器性能的影响.常州工学院学报.2003年第16卷第2期.
[6] RFC2068: Hypertext Transfer Protocol-HTTP/1.1, Networking Group, Jan 1997.
[7] 鲍剑洋,沈群.Web服务器性能优化.计算机工程与应用.1999年第6期.
[8] 武海平,蒋东兴,程志锐,康晓宁.Web服务器通用性能测试系统的设计与实现.小型微型计算机系统.2003年第24卷第2期.
[9] WebStone: The First Generation in HTTP Server Benchmarking, Gene Trent, Feb 1995.
[10] SpecWeb99: Standard Performance Evaluation Coporation, June 1999.
[11] 韩光涛,张元英.SpecWeb99:最新Web服务器性能基准测试.计算机应用研究.2000年第4期.
[12] 陈新中,李岩,杨炳儒,谢永红.Web日志挖掘技术进展.系统工程与电子技术.2003年第25卷第4期.
[13] SRIVASTAVA J, COOLEY R, DESHPANDA M, et al. Web usage mining: discovered applications of usage pattern from Web data [J]. ACM SIGKDD Exploration, 2000, 1(2):12-23.
[14] 王实,高文,李锦涛.Web数据挖掘[J].计算机科学,2000,27(4):28-31.
[15] COOLEY R, MOBASHER B, SRIVASTAVA J.Web mining: information and pattern discovery on the World Wide Web [Z].Internatinal conference on tools with aritificial intelligence.newport, neach, Canada, 1997.
[16] COOLEY R, MOBASHER B, SRIVASTAVA J.Data preparation for mining World Wide Web browsing patterns [J]. Journal of Knowledge Information Systems, 1999, 1(1):5-32.
[17] PIROLLIP, PITKOWJ, RAOR. Silk from a Sow's ear: extracting usable structures from the Web [Z].1996 Conferences on Human factors in computing systems (CHI-96), VancouVer, British Columbia, Canada, 1996.
[18] AGRAWL R, SRIKANT R.Fast algorithm for mining association rule [Z].20th VLDB Conference Santiage, Chile, 1994.
[19] HAN J, KAMBR N.Data mining concept and techniques [Z].Bei Jing: High Education Press, 1999.
[20] SAVASERE A, OMIEINSKI E, SNAVATHE.An efficient algorithm from mining association rules [Z].21th VLDB Conference, Zurich, Switzerland, 1995.
[21] GOPALAN R, GOPLAN NAIR M.Discovering fuzzy association rules in large data relation [Z].The 9th Australian: Database Conference, ADC98 PERTH, Australian, 1998.
[22] AGRAWAL R, SRIKAKANT R.Mining Sequential pattern[Z].1995 Int Conf Data Engineering,Taipei,Taiwan,1995.,
[23] AGRAWAL R, SRIKAKANT R.Mining quantitive association rules in large relation tables [Z].1996 ACM-SIGMOD Conf Management of Data, Montreal, Canada, 1996.
[24] PEI J, HAN J W.Mining access pattern efficiently from Web Mining Logs.[J/OL].4th pacific-Asia conference Kyoto, Japan, 2000.
[25] SPILIOPOULOU M.The Laborious way from data mining to Web Log Mining [J]. Computer System Science & Engineer 1999, 14(2):113-126.
[26] 陆丽娜,魏恒义,杨怡玲等.Web日志挖掘中的序列模式识别[J].小型微型计算机系统,2000,21(5):481-483.
[27] 叶施仁,史忠植,吴斌等.KDD中的机器学习[J].计算机科学,2001,28(3):96-100.
[28] 王实,高文.数据挖掘中的聚类方法[J].计算机科学,2000,27(4):42-45.
[29] ZHANG T, RAMAKRISHNAN R, LIVNY M. BIRCH: An efficient data clustering method for very large database [Z].1996 ACM-SIGMOD Int Conf Management of Data, Montreal, Canada, 1996.
[30] NASRAOUI O, FRIGUI H,JOSHI A.Mining Web Access Logs using relational competitive fuzzy clustering[Z].The Eight International Fuzzy System Association World Congress, Taipei, Taiwan, 1999.
[31] FU Y, SANDHU K, SHIH M, Clustering of Web user based on access pattern[Z].The 1999 KDD Workshop on Web Mining,San Diego,Canada,1999.
[32] YEUNG DS, WANG XZ.Improving performance of similiarity based clustering by feature weight learning [J].IEEE Transactions on pattern Analysis and Machine Intelligence, 2002, 24(4):556-561.
[33] DUDA T, HART P. Pattern classification and Science analysis [Z].John Wiley and Sons, New York, 1973.
[34] 王实,高文.增强型朴素贝叶斯学习[J].计算机科学,2000,27(4):46-49.
[35] 欧阳为民,郑诚,张燕.知识发现与数据挖掘工具评述[J].计算机科学,2001,28(3):101-108.
[36] QUIN LAN J R. Induction of decision tree [J]. Machine Learning, 1986, 1:81-106.
[37] QUIN LAN J R.C4.5: Programs for machine learning [Z].Morgan Kaufmann, San Master, Canada, 1993.
[38] ZAI ANE O,MAX,JHAN.Discovering Web access patterns and trends by applying OLAP and data mining technology on Web Logs[Z].Advances in Digital Libraries,Santa Barbara, Canada,1998.
[39] PITKMOW J, KRISHNA K, and BHARAT.Webviz: A tool for World Wide Web access log Analysis [Z].First International WWW Conference, Geneva, Switzerland, 1994.
[40] MOBASHER B,JAIN N,HAN E,et al.Pattern discovery from world wide web transaction[Z].Technical Report TR 96050,University of Minnesota,1996.
[41] 赵晓芳,刘欣,徐志伟.网络交通自相似特性的分析及应用——具有单一登录点的机群网络服务器的性能评测.计算机研究与发展,Vol.36,No.9,Sep.1999.
[42] Walter W, Murad S T, Robert S, etal. Self-similarity through high-variability: Statistical analysis of Ethernet L AN traffic at the sourceLevel. IEEE/ACM Trans on Networking, 1997,5(1):71～86.
[43] Vern P, Sally F. Wide-area traffic: The failure of possion modeling. IEEE/ACM Trans on Networking, 1995, 3(3): 226～244.
[44] L eland W E, OttT T.UNIX process behavior and load balancing among loosely-coupled computers.Teletraffic Analysis and Computer Perform ance Evaluation. Amsterdam: Elsevier Science Publishers B U, 1986. 191～208.
[45] E. Fuchs and P.E.Jackson, "Estimates of Distributions of Random Variables for Certain Computer Communication Traffic Models", Comm. of ACM, Vol. 13, No 12, December 1970, pages 752-767.
[46] D.E. Duffy, A.A.Mcintosh, M. Rosenstain and W. Willinger, "Statistical analysis fo CCSN/SS7 Traffic Data from Working CCS Subnetworks", IEEE Journal of Selected Areas in Communication, Vo1.12, No. 3, April 1994, Pages 544-551.
[47] W.E. Leland, M.S. Taqqu, W. Willinger and D.V. Wilson, "On the Self-Similar Nature of Ethernet Traffic", IEEE/ACM Transactions on Networking, Vol.2, No.l, February 1994, pages 1-15.
[48] Klivansky, S. K., Mukherjee,A. and Song, C. "On Long Range Dependence in NSFNET Traffic" Technical Report GIT-CC-94/61, Geogia Institute of Technology, Altlanta, GA 30332, USA, December 1994.
[49] R. Jain, "ATM Networking: Issues and Challenges Ahead", Networld+InterOp'95 Engineer Conference, Las Vegas, Nevada, March 27-31, 1995.
[50] W L etand, M Taqqu, W Willinger et al. On the self-similar nature of Ethernet traffic (extended version).IEEE/ACM Trans on Networking, 1994, 2(1): 1～15.
[51] W Willinger, M Taqqu, R Sherm an et al. Self-similarity through high-variability: Statistical analysis of Ethernet L AN traffic at the source level. IEEE/ACM Trans on Networking, 1997, 5(1): 71～86.
[52] V Paxson. Empirically derived analytic models of wide-area TCP connections. IEEE/ACM Trans on Networking, 1994, 2(4): 316～336.
[53] B Tsybakov, N Georganas. On self-similar traffic in ATM queues: Definitions, overflow probability bound, and cell delay distribution. IEEE/ACM Trans on Networking, 1997,5 (3): 339～346.
[54] M Crovella, A Bestavros. Self-similarity in World Wide Web traffic evidence and possible causes. In: SIGMETRICS'96. Marriott, PA, 1996.
[55] M Crovella, A Bestavros. Explaining World Wide Web traffic self-similarity. Department of Computer Science of Boston University, Tech Rep: TR-95-015, 1995.
[56] 肖志新,杨岳湘,杨霖,吴利华.网络流量的自相似特性.吉首大学学报(自然科学版),第25卷,第2期,2004年6月.
[57] Paul Barford et al, Generating Representative Web WorkLoads for Network and Server Performance Evaluation.Technical Report Bu-CS-97-006,Boston University Computer Science Department,December 1997.
[58] W Stallings. High-Speed Networks TCIP/IP and ATM Design Principles. Englewood Cliffs, NJ: Prentice Hall, 1998.
[59] C Cunha, A Bestavros, M Crovella. Characteristics of WWW client-based traces. Department of Computer Science of Boston University, Tech Rep: TR-95-010, 1995.
[60] M Arlitt, C Williamson. Web server workload characterization: The search for invariants (extended version). In: ACMSIGMETRICS' 96. Marriott, PA, 1996.
[61] DROZ P, BOUDEC J L. A high_speed self_similar ATM VBR traffic generator [J].IEEE Global Telecommunication Conference, 1996(11).
[62] PAXSON V. Fast, approximate synthesis of fractional Gaussian noise for generating self_similiar network traffic [J].Computer Communications Review, 1997(10):5-18.
[63] CROVELLA M K, BESTAVROS A.Self-similarity in World Wide Web traffic evidence and possible causes [J].IEEE/ACM Transactions on Networking, 1997, 5(6):835-846.
[64] PARK K,KIM G,CROVELLA M K.On the relationship between file sizes,transport protocols,and self-similar network traffic[J].IEEE International Conference on Network Protocols, 1996:171-180.
[65] TUAN T, PARK K.Multiple time scale redundancy control for QoS_sensitive transport of real time traffic [J].IEEE INFOCOM, 2000, Tel_Aviv, Israel.
[66] NORROS I.On the use of fractional Brownian motion in the theory of connectionless networks [J].IEEE JSAC, 1995, 13(6):953-962.
[2] Zona Research Inc. The economic impacts of unacceptable web-site downloadspeed. 1999.http://www.keynote.com/downloads/whitepapers/economic_impact_of_downlaodspeed.pdf.
[3] ARudolf, RPirker. E-Business testing: User perceptions and performance issues.In: Proc of the 1st Asia Pacific Conf on Quality Software. LosAlamitos: IEEE Computer Society Press, 2000 315～323.
[4] SPECweb99 Release Documents 1.02 July 21,.2000. http://open.spec.org/web99/docs/whitepaper.html
[5] 查志琴,李慧.HTTP协议的发展对Web服务器性能的影响.常州工学院学报.2003年第16卷第2期.
[6] RFC2068: Hypertext Transfer Protocol-HTTP/1.1, Networking Group, Jan 1997.
[7] 鲍剑洋,沈群.Web服务器性能优化.计算机工程与应用.1999年第6期.
[8] 武海平,蒋东兴,程志锐,康晓宁.Web服务器通用性能测试系统的设计与实现.小型微型计算机系统.2003年第24卷第2期.
[9] WebStone: The First Generation in HTTP Server Benchmarking, Gene Trent, Feb 1995.
[10] SpecWeb99: Standard Performance Evaluation Coporation, June 1999.
[11] 韩光涛,张元英.SpecWeb99:最新Web服务器性能基准测试.计算机应用研究.2000年第4期.
[12] 陈新中,李岩,杨炳儒,谢永红.Web日志挖掘技术进展.系统工程与电子技术.2003年第25卷第4期.
[13] SRIVASTAVA J, COOLEY R, DESHPANDA M, et al. Web usage mining: discovered applications of usage pattern from Web data [J]. ACM SIGKDD Exploration, 2000, 1(2):12-23.
[14] 王实,高文,李锦涛.Web数据挖掘[J].计算机科学,2000,27(4):28-31.
[15] COOLEY R, MOBASHER B, SRIVASTAVA J.Web mining: information and pattern discovery on the World Wide Web [Z].Internatinal conference on tools with aritificial intelligence.newport, neach, Canada, 1997.
[16] COOLEY R, MOBASHER B, SRIVASTAVA J.Data preparation for mining World Wide Web browsing patterns [J]. Journal of Knowledge Information Systems, 1999, 1(1):5-32.
[17] PIROLLIP, PITKOWJ, RAOR. Silk from a Sow's ear: extracting usable structures from the Web [Z].1996 Conferences on Human factors in computing systems (CHI-96), VancouVer, British Columbia, Canada, 1996.
[18] AGRAWL R, SRIKANT R.Fast algorithm for mining association rule [Z].20th VLDB Conference Santiage, Chile, 1994.
[19] HAN J, KAMBR N.Data mining concept and techniques [Z].Bei Jing: High Education Press, 1999.
[20] SAVASERE A, OMIEINSKI E, SNAVATHE.An efficient algorithm from mining association rules [Z].21th VLDB Conference, Zurich, Switzerland, 1995.
[21] GOPALAN R, GOPLAN NAIR M.Discovering fuzzy association rules in large data relation [Z].The 9th Australian: Database Conference, ADC98 PERTH, Australian, 1998.
[22] AGRAWAL R, SRIKAKANT R.Mining Sequential pattern[Z].1995 Int Conf Data Engineering,Taipei,Taiwan,1995.,
[23] AGRAWAL R, SRIKAKANT R.Mining quantitive association rules in large relation tables [Z].1996 ACM-SIGMOD Conf Management of Data, Montreal, Canada, 1996.
[24] PEI J, HAN J W.Mining access pattern efficiently from Web Mining Logs.[J/OL].4th pacific-Asia conference Kyoto, Japan, 2000.
[25] SPILIOPOULOU M.The Laborious way from data mining to Web Log Mining [J]. Computer System Science & Engineer 1999, 14(2):113-126.
[26] 陆丽娜,魏恒义,杨怡玲等.Web日志挖掘中的序列模式识别[J].小型微型计算机系统,2000,21(5):481-483.
[27] 叶施仁,史忠植,吴斌等.KDD中的机器学习[J].计算机科学,2001,28(3):96-100.
[28] 王实,高文.数据挖掘中的聚类方法[J].计算机科学,2000,27(4):42-45.
[29] ZHANG T, RAMAKRISHNAN R, LIVNY M. BIRCH: An efficient data clustering method for very large database [Z].1996 ACM-SIGMOD Int Conf Management of Data, Montreal, Canada, 1996.
[30] NASRAOUI O, FRIGUI H,JOSHI A.Mining Web Access Logs using relational competitive fuzzy clustering[Z].The Eight International Fuzzy System Association World Congress, Taipei, Taiwan, 1999.
[31] FU Y, SANDHU K, SHIH M, Clustering of Web user based on access pattern[Z].The 1999 KDD Workshop on Web Mining,San Diego,Canada,1999.
[32] YEUNG DS, WANG XZ.Improving performance of similiarity based clustering by feature weight learning [J].IEEE Transactions on pattern Analysis and Machine Intelligence, 2002, 24(4):556-561.
[33] DUDA T, HART P. Pattern classification and Science analysis [Z].John Wiley and Sons, New York, 1973.
[34] 王实,高文.增强型朴素贝叶斯学习[J].计算机科学,2000,27(4):46-49.
[35] 欧阳为民,郑诚,张燕.知识发现与数据挖掘工具评述[J].计算机科学,2001,28(3):101-108.
[36] QUIN LAN J R. Induction of decision tree [J]. Machine Learning, 1986, 1:81-106.
[37] QUIN LAN J R.C4.5: Programs for machine learning [Z].Morgan Kaufmann, San Master, Canada, 1993.
[38] ZAI ANE O,MAX,JHAN.Discovering Web access patterns and trends by applying OLAP and data mining technology on Web Logs[Z].Advances in Digital Libraries,Santa Barbara, Canada,1998.
[39] PITKMOW J, KRISHNA K, and BHARAT.Webviz: A tool for World Wide Web access log Analysis [Z].First International WWW Conference, Geneva, Switzerland, 1994.
[40] MOBASHER B,JAIN N,HAN E,et al.Pattern discovery from world wide web transaction[Z].Technical Report TR 96050,University of Minnesota,1996.
[41] 赵晓芳,刘欣,徐志伟.网络交通自相似特性的分析及应用——具有单一登录点的机群网络服务器的性能评测.计算机研究与发展,Vol.36,No.9,Sep.1999.
[42] Walter W, Murad S T, Robert S, etal. Self-similarity through high-variability: Statistical analysis of Ethernet L AN traffic at the sourceLevel. IEEE/ACM Trans on Networking, 1997,5(1):71～86.
[43] Vern P, Sally F. Wide-area traffic: The failure of possion modeling. IEEE/ACM Trans on Networking, 1995, 3(3): 226～244.
[44] L eland W E, OttT T.UNIX process behavior and load balancing among loosely-coupled computers.Teletraffic Analysis and Computer Perform ance Evaluation. Amsterdam: Elsevier Science Publishers B U, 1986. 191～208.
[45] E. Fuchs and P.E.Jackson, "Estimates of Distributions of Random Variables for Certain Computer Communication Traffic Models", Comm. of ACM, Vol. 13, No 12, December 1970, pages 752-767.
[46] D.E. Duffy, A.A.Mcintosh, M. Rosenstain and W. Willinger, "Statistical analysis fo CCSN/SS7 Traffic Data from Working CCS Subnetworks", IEEE Journal of Selected Areas in Communication, Vo1.12, No. 3, April 1994, Pages 544-551.
[47] W.E. Leland, M.S. Taqqu, W. Willinger and D.V. Wilson, "On the Self-Similar Nature of Ethernet Traffic", IEEE/ACM Transactions on Networking, Vol.2, No.l, February 1994, pages 1-15.
[48] Klivansky, S. K., Mukherjee,A. and Song, C. "On Long Range Dependence in NSFNET Traffic" Technical Report GIT-CC-94/61, Geogia Institute of Technology, Altlanta, GA 30332, USA, December 1994.
[49] R. Jain, "ATM Networking: Issues and Challenges Ahead", Networld+InterOp'95 Engineer Conference, Las Vegas, Nevada, March 27-31, 1995.
[50] W L etand, M Taqqu, W Willinger et al. On the self-similar nature of Ethernet traffic (extended version).IEEE/ACM Trans on Networking, 1994, 2(1): 1～15.
[51] W Willinger, M Taqqu, R Sherm an et al. Self-similarity through high-variability: Statistical analysis of Ethernet L AN traffic at the source level. IEEE/ACM Trans on Networking, 1997, 5(1): 71～86.
[52] V Paxson. Empirically derived analytic models of wide-area TCP connections. IEEE/ACM Trans on Networking, 1994, 2(4): 316～336.
[53] B Tsybakov, N Georganas. On self-similar traffic in ATM queues: Definitions, overflow probability bound, and cell delay distribution. IEEE/ACM Trans on Networking, 1997,5 (3): 339～346.
[54] M Crovella, A Bestavros. Self-similarity in World Wide Web traffic evidence and possible causes. In: SIGMETRICS'96. Marriott, PA, 1996.
[55] M Crovella, A Bestavros. Explaining World Wide Web traffic self-similarity. Department of Computer Science of Boston University, Tech Rep: TR-95-015, 1995.
[56] 肖志新,杨岳湘,杨霖,吴利华.网络流量的自相似特性.吉首大学学报(自然科学版),第25卷,第2期,2004年6月.
[57] Paul Barford et al, Generating Representative Web WorkLoads for Network and Server Performance Evaluation.Technical Report Bu-CS-97-006,Boston University Computer Science Department,December 1997.
[58] W Stallings. High-Speed Networks TCIP/IP and ATM Design Principles. Englewood Cliffs, NJ: Prentice Hall, 1998.
[59] C Cunha, A Bestavros, M Crovella. Characteristics of WWW client-based traces. Department of Computer Science of Boston University, Tech Rep: TR-95-010, 1995.
[60] M Arlitt, C Williamson. Web server workload characterization: The search for invariants (extended version). In: ACMSIGMETRICS' 96. Marriott, PA, 1996.
[61] DROZ P, BOUDEC J L. A high_speed self_similar ATM VBR traffic generator [J].IEEE Global Telecommunication Conference, 1996(11).
[62] PAXSON V. Fast, approximate synthesis of fractional Gaussian noise for generating self_similiar network traffic [J].Computer Communications Review, 1997(10):5-18.
[63] CROVELLA M K, BESTAVROS A.Self-similarity in World Wide Web traffic evidence and possible causes [J].IEEE/ACM Transactions on Networking, 1997, 5(6):835-846.
[64] PARK K,KIM G,CROVELLA M K.On the relationship between file sizes,transport protocols,and self-similar network traffic[J].IEEE International Conference on Network Protocols, 1996:171-180.
[65] TUAN T, PARK K.Multiple time scale redundancy control for QoS_sensitive transport of real time traffic [J].IEEE INFOCOM, 2000, Tel_Aviv, Israel.
[66] NORROS I.On the use of fractional Brownian motion in the theory of connectionless networks [J].IEEE JSAC, 1995, 13(6):953-962.
© 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 |