成功大學博碩士論文系統

簡易檢索 / 詳目顯示

回結果列表

研究生：	紀勝中 -Chung, Sheng
論文名稱：	智慧型網路效能評估器 Intelligent Network Performance Evaluator
指導教授：	鄭芳田 Cheng, Fan-Tien 楊浩青 Yang, H.-C.
學位類別：	碩士 Master
系所名稱：	電機資訊學院 - 製造工程研究所 Institute of Manufacturing Engineering
論文出版年：	2007
畢業學年度：	95
語文別：	中文
論文頁數：	60
中文關鍵詞：	網路可用度、效能指標、曲線擬合
外文關鍵詞：	Curve Fitting, Performance Index, Network Availability
相關次數：	點閱：139 下載：0
分享至:	分享至facebook 分享至twitter

查詢本校圖書館目錄查詢臺灣博碩士論文知識加值系統勘誤回報

對製造業而言，具有高可用度的網路環境是相當重要的，因網路失
效可能會造成生產停擺，致產生極大的成本損失。傳統的網路效能管理著重於監控網路的流量，以偵測網路負載的變化。但此種方法僅能監控網路，卻無法偵測出因網路負載過重時，可能引起的網路性能下降致網路失效之情形。
本論文提出一個即時監控計算機關鍵資源之智慧型網路效能評估器
(Intelligent Network Performance Evaluator, INPE)。INPE 具有失效偵測及預測失效時間等功能。失效偵測模組將應用效能指標 (Performance Index)來判斷節點的效能狀態。當偵測出該節點已有性能衰退之現象後，將通知預測模組去預測失效剩餘時間(Time to Failure)。失效預測模組將會根據該節點剩餘的可用資源來預測失效剩餘時間。本論文利用曲線擬合法來估算失效剩餘時間，並提出一個效能改進策略，以延長網路的可用時間。
INPE 除了可應用來監測一般的流量外，也可以偵測出網路衰退情形
並預測其失效剩餘時間。在網路負戴有65~90%的情況下，可運用效能改進策略來延長網路的使用時間：在預測失效時間剩30 分鐘時，藉由每阻檔10 分鐘，可讓網路的可利用時間增長1.5 小時至2 小時。

High availability networking environment is essential for a modern production factory.Network failures may shut production lines down. To evaluate network performance, conventional methods focused on monitoring the overall network flows. However, those methods may not be able to discover the major cause of network performance degradation. This major cause may be due to over-loading of a node.
To avoid service loss due to network failures, this work proposes an Intelligent Network Performance Evaluator (INPE) with two kernel modules: failure detection and time-to-failure (TTF) prediction. In the failure detection module, instead of flow monitoring, INPE uses principle component analysis of the resource counters to construct the performance index for determining the network performance of a node. When the index indicates degradation of a node, the prediction module will be triggered to predict the TTF of the node’s networking. Using linear-regression or curve-fitting methods, the TTF module assures the predicted TTF is in an acceptable range. This work also proposes a blocking strategy to extend the available time of networking.
In a network environment with 60%~95% loading, the experimental results show that the mean accuracy of prediction is over 78% by the curve-fitting method. In this case, the available time of network may be extended for 1.5~2 hours by the blocking strategy when the remaining TTF is 30 minutes and the over-loading node is blocked for 10 minutes.

中文摘要
英文摘要
致謝
目錄....................................................................................................................i
圖目錄............................................................................................................... v
表目錄............................................................................................................viii
第一章 緒論
1 背景.................................................................................................. 1
2 研究動機與目的.............................................................................. 2
3 研究範圍與限制.............................................................................. 3
4 論文架構.......................................................................................... 3
第二章 文獻探討與基礎理論
1 相關文獻探討.................................................................................. 4
1.1 網路負載量測........................................................................... 4
1.2 網路負載分析........................................................................... 4
1.3 失效偵測分析........................................................................... 5
2 基礎理論.......................................................................................... 6
2.1 相關分析................................................................................... 6
2.2 主成份分析............................................................................... 6
2.3 Lagrange Polynomial ............................................................... 11
2.4 簡單線性迴歸......................................................................... 11
2.5 曲線擬合................................................................................. 12
3 網路失效分析................................................................................ 14
3.1 資源計數器............................................................................. 14
3.2 簡易網路管理協定................................................................. 16
3.2.1 網路的管理與控制......................................................... 17
3.2.2 管理者與代理者............................................................. 17
3.2.3 管理資訊的結構與管理資訊庫..................................... 19
3.2.4 UDP................................................................................... 19
第三章 系統架構設計
1 實驗環境........................................................................................ 21
2 INPE 之軟體架構.......................................................................... 22
2.1 偵測模組................................................................................. 23
2.2 預測模組................................................................................. 23
3 智慧型網路效能評估器流程圖................................................... 24
3.1 偵測模組流程圖..................................................................... 24
3.2 預測模組流程圖..................................................................... 25
4 效能指標演算法............................................................................ 28
第四章 實驗結果與比較
1 評估網路效能之實驗環境............................................................ 34
1.1 實驗合理性驗證...................................................................... 34
1.2 在不同網路利用率下的實驗時間......................................... 35
1.3 局部化之網路失效.................................................................. 39
2 在網路利用率90%下之實驗結果分析....................................... 39
3 偵測模組門檻值設定.................................................................... 41
3.1 快速傅利葉轉換..................................................................... 41
3.2 連續 Timeout 與效能指標關聯性分析............................... 44
3.3 效能指標成份......................................................................... 44
3.4 偵測模組之效能指標門檻值................................................ 45
4 預測模組實驗結果分析................................................................ 47
4.1 方法比較................................................................................. 47
4.2 模型驗證................................................................................. 50
5 效能改善方案................................................................................ 52
5.1 Block 流程圖.......................................................................... 53
5.2 Block 實驗結果分析.............................................................. 54
第五章 結論與未來展望
1 研究成果與論文總結.................................................................... 57
2 未來研究方向................................................................................ 57
參考文獻............................................................................................... 59
                                    

[1] C. Jungyun, “Introduction to Semiconductor Manufacturing”, A Special Session for ICRA 2001: Automation in Semiconductor Industry.
[2] 宏碁新聞稿. http://global.acer.com/t_chinese/about/news.asp?id=6284
[3] 陳秋彰「在無線區域網路上之位置導向效能管理」，國立暨南國際大學資訊管理研究所碩士論文，2004。
[4] 陳逸「Sniffer Pro 網路最佳化與故障排除參考手冊」，博碩，2005。
[5] F.-T. Cheng, H.-C. Yang, and C.-Y. Tsai, “Developing a Service Management
Scheme for Semiconductor Factory Management Systems,” IEEE Robotics and
Automation Management, vol. 11, no. 1, pp. 26-40, March 2004.
[6] F.-T. Cheng, S.-L. Wu, P.-Y. Tsai, Y.-T. Chung, and H.-C. Yang, “Application
Cluster Service Scheme for Near-Zero-Downtime Services,” in Proc. 2005 IEEE
Conference on Robotics and Automation, pp. 4062 - 4067, Apr. 2005.
[7] 涂哲源「建構在 ARM 平台上的 IPE」，國立成功大學製造工程研究所碩士論文，2006。
[8] K. Lai and B. Mary, “Measuring Bandwidth,” Proceedings of IEEE INFOCOM,
vol.1, pp.235-245, 1999.
[9] M. Liu, J. Shi, Z. Li, Z. Kan, and J. Ma, “A new end-to-end measurement method
for estimating available bandwidth,” IEEE Symposium on Computers and
Communications, vol.2, pp.1393-1400, 2003.
[10] 江巧雯「應用時間序列叢集技術於網路流量分級之研究」，元智大學資訊管理研究所碩士論文，2000。
[11] H. H. Yue and M. Tomoyasu, “Weighted Principal Component Analysis and its
Applications to Improve FDC Performance,” IEEE Conference on Decision and
Control, pp. 4262-4267, Dec. 2004.
[12] S.-S. Kim and J.-L. Crompton, “The effects of different types of information
messages on perceptions of price and stated willingness-to-pay,” Journal of Leisure
Research, vol.3, pp. 299-318, 2001.
[13] R. A. Johnson and D. W. Wichern, “Applied Multivariate Statistical Analysis,”
Prentice-Hall, 2002.
[14] 陳順宇「多變量分析」，華泰書局，2004。
[15] 陳順宇「統計學學」，華泰書局，2004。
[16] Sauer and Tim, “Numerical Analysis,” Addison-Wesley, 2005.
[17] A. Shamir, “How to share a secret,” Communication of the ACM, vol. 22, Issue 11,pp. 612-613, 1979.
[18] A. T. Curns and M. Azhar, “Regression Analysis and Other Multivariable Methods,”Duxbury Press, 1998.
[19] 張智星「MATLAB 程式設計」，清蔚科技，2004。
[20] Microsoft Management Console.
http://technet2.microsoft.com/WindowsServer/zh-CHT/Library/329ce1bd-9bb4-4b6
3-947e-0d1e993dc27d1028.mspx?mfr=true.
[21] Microsoft TechNet Monitoring and Troubleshooting Performance.
http://technet.microsoft.com/en-us/default.aspx
[22] 陳麗媛「在 AIS 平台上實現網際網路流量」，國立中正大學電機工程研究所碩士論文，1999。
[23] J. Case, M. Fedor, M. Schoffstall, and J. Davin, “A Simple Network Management
Protocol”, Internet Engineering Task Force working note, Network Information
Center, SRI International, Menlo Park, California, 1988.
[24] STALLINGS, “SNMP, SNMPV2, AND CMIP,” Addison-Wesley, 1993.
[25] K. McCloghrie and M. Rose, “Management Information Base for Network
Management of TCP/IP-based internets: MIB II,” RFC 1213, IETF, 1991.
[26] 劉琍綾「SNMP 網管實務」，歐萊禮，2003。
[27] 英文調查報告書 “U.S. Location-based Service (LBS) Markets,” Frost & Sullivan,2005.

校內：2107-09-03公開
校外：2107-09-03公開電子論文尚未授權公開，紙本請查館藏目錄

簡易檢索 / 詳目顯示

相關論文