臺灣博碩士論文加值系統

隨著網際網路的重要性不斷地提升，網路上的各種應用服務之需求亦隨之增加。網路內容供應商面對許多使用者對於網路內容的大量需求。不過，單一的網頁伺服器並無法回應如此大量的需求。因此，第四層伺服器叢集系統成為目前最主要的網頁伺服器系統。然而，第四層交換器的負載平衡只能看到封包的傳輸層以下的欄位，而沒辦法針對封包裡面的內容作處理，換言之，第四層交換器的負載平衡並沒辦法支援目前網路上最重要的服務，電子商務。也因此，許多廠商推出了網頁交換器這類支援網路應用層的交換器，支援各類較複雜的網路服務。
本研究提出了一個針對第七層網路設備的測試系統。換言之，就是針對目前市面上網頁交換器此類產品的功能，設計了一套測試平台。此系統定義了一套XML語言來描述網路使用者的行為，目前主要在網路上使用者的主要行為包括了網頁的瀏覽跟電子商務的交易，換言之，網頁的瀏覽代表了靜態網頁而電子商務代表動態網頁。目前的一些針對網頁交換器的測試工具主要著重在測試負載平衡處理的功能，測量的單位是每秒能處理多少HTTP要求的能力。此外，本研究提出了關於測試持續性（persistence）功能的觀念。持續性（persistence）可以完整地支援電子商務的交易，使其在一個伺服器叢集系統的環境下運作而不產生交易上的錯誤。本研究所提出的測試系統可以測量網頁交換器處理電子商務的效能與正確性。另外，藉由XML的設計，本系統也維持了高度的可擴充性。

With the growing importance of World Wide Web, the need of web service is growing fast. Many Internet content provider faces huge demand of users. A single Web server cannot provide such a huge response. So a Layer-4 web server cluster is currently the main web server system. However, L4 switches are optimized for the transport layer, they are completely unaware of the Application Layer, which in the Internet, is HTTP (L5 or L5-L7). That means Layer-4 Load Balancing cannot support E-commerce, which is the popular demand in the Internet. Therefore, the Layer-7 Web server cluster is provided to satisfy the complex demand in the Internet.
This thesis presents a testing system for Layer-7 network devices. Since the main product of Layer-7 network devices is Web Switch, this thesis aims the target of testing Web Switch technology. This system defines a XML language to depict user’s behavior. Major behavior of people in the Internet is browsing and E-commerce. In other words, browsing means getting static pages, and E-commerce means getting dynamic pages. Current testing tools for the Web Switch mainly measures the performance by getting static pages with URL load balancing; the unit is request/second. This thesis presents a testing platform to measure another testing item to measure the Web Switch; that is persistence. Because persistence offered by the Web Switch can completely support e-commerce without transaction error. This system can measure the performance and correctness of E-commerce supported by the Web Switch. With XML configuration, this system also provides extensibility. Extensibility provides testing of other Layer-7 network devices in the future.

Chapter 1. Introduction 1
Chapter 2. Layer-7 Network Devices Technology 4
2.1. Routing 6
2.1.1. URL-Based Server Load Balancing 6
2.1.2. Cookie-Based Preferential Load Balancing 7
2.2. Persistence 7
2.2.1. Cookie persistence 7
2.2.2. SSL persistence 10
Chapter 3. Recent Testing Tool 12
3.1. Limitation of current tool 13
Chapter 4. Testing Platform Implementation 15
4.1. User Behavior Language Definition 15
4.1.1. Design of “Client.xml” 16
4.1.2. Design of “flow.xml” 21
4.2. System architecture 23
4.2.1. Operation 26
4.2.2. Controller 28
4.2.3. Traffic generator 29
Chapter 5. Experiment 33
5.1. Experiment setup 33
5.2. Experiment Result 34
Chapter 6. Conclusion 37
Chapter 7. Reference 38

[1.] George Apostolopoulos, David Aubespin, Vinod Peris, Prashant Pradhan, Debanjan Saha, “Design, Implementation and Performance of a Content-Based Switch”, INFOCOM 2000. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies. Proceedings. IEEE , Volume: 3 , 2000 , Page(s): 1117 -1126 vol.3
[2.] Zornitza Genova and Kenneth J. Christensen, “Challenges in URL Switching for Implementing Globally Distributed Web Sites”, Parallel Processing, 2000. Proceedings. 2000 International Workshops on , 2000,
Page(s): 89 —94
[3.] K.Y. Leung, K.H. Yeung, “The Design and Implementation of a WWW Traffic Generator”, Parallel and Distributed Systems, 2000. Proceedings. Seventh International Conference on , 2000, Page(s): 509 -514
[4.] D. Kristol, Bell Laboratories, Lucent Technologies, L. Montulli, Epinions.com, ” HTTP State Management Mechanism”, October 2000, rfc 2965
[5.] R. Fielding, UC Irvine, J. Gettys , …, “Hypertext Transfer Protocol — HTTP/1.1”, June 1999, rfc 2616
[6.] S. Bradner, Harvard University, J. McQuaid, NetScout Systems, “Benchmarking Terminology for Network Interconnection Devices”, March 1999, rfc 2544
[7.] V. Pai, M. Aron, M. Svendsen, G. Banga, P. Druschel, W. Zwaenepoel, and E. Nahum, “Locality-aware request distribution in cluster-based network servers”, Proceedings of the 8th International Conference on Architectural Support for Programming Languages and Operation Systems, October 1998
[8.] Web Polygraph, http://www.web-polygraph.org/
[9.] The Measurement Factory Inc., http://www.measurement-factory.com/
[10.] WebBench, http://www.etestinglabs.com/benchmarks/webbench/webbench.asp
[11.] SPECWeb99 Benchmark, http://www.specbench.org/osg/web99/
[12.] WebSTONE, http://www.mindcraft.com/webstone/
[13.] Extensible Markup Language (XML)1.0, W3C Recommendation 6 October 2000, http://www.w3.org/TR/2000/REC-xml-20001006/
[14.] Tolly Group, “Alteon versus ArrowPoint”, http://wp.bitpipe.com/resource/org_905740864_63/TollyTS200204AlteonandArrowPointL7Mar00.pdf
[15.] PC Magazine Labs, “Alteon versus ArrowPoint”, http://www.zdnet.co.uk/pcmag/ne/shoot-out/2000/04/10.html
[16.] University of Wisconsin Real-World Labs, “F5 versus Alteon versus Foundry versus Intel”, http://www.networkcomputing.com/1115/1115f1.html
[17.] Alteon, Web OS 9.0 Application Guide(050159A)
[18.] 林柏青, 蔡品再, 林盈達, “網頁交換器產品評比─功能與效能面”, 網路通訊雜誌
[19.] Linux, http://www.linux.org/
[20.] PHP, http://www.php.net
[21.] Apache, http://www.apache.org
[22.] MSDN, http://msdn.microsoft.com/