臺灣博碩士論文加值系統

　　多元件模組化冗餘(NMR)技術已被廣泛地用來提昇硬體系統的可靠度。同樣地，軟體版本多樣化冗餘(NVDR)策略已被宣稱可用以提昇軟體可靠度。在主從式資料庫管理系統，可靠度的焦點在於能夠正確無誤地執行運作。因此，不論是軟體或是硬體的可靠度皆會影響資料在系統中處理的效率與精確性。目前在硬體系統中已有完整的可靠度模式來定義、評估及提昇系統可靠度。不幸地，在軟體結構方面，僅只有用以提昇軟體可靠度的策略－NVDR，並無一個明確且完整的系統表示法來正確地表達特定的軟體結構，因此，系統架構(framework)表示法是必需發展的。本論文提出利用系統架構與可靠度分析方法以有效地採用多元件模組化冗餘技術及軟體版本多樣化冗餘策略來建立較可靠的資料庫系統。
本論文主要的目的在根據會直接影響資料庫系統運作之可靠度的主要效用因子，來明確地說明主從式資料庫系統的實體架構。並且提出一個完整的及系統化的序列式系統架構表示法，以期能以簡明的及正確無誤的系統定義來描述特定的主從式資料庫系統。為了能夠迅速且無誤地評估資料庫系統的可靠度，本論文亦依據主從式資料庫系統中用戶端、連接網路、伺服端之系統特性整理出系統評估標準並建立系統可靠度索引表。使用序列式系統架構及系統可靠度索引表可以正確地顯示經由同時應用多元件模組化冗餘技術及軟體版本多樣化冗餘策略來評估組合了硬體及軟體的系統可靠度，並且可以明確地指出在限定的資源下，應用何種冗餘的系統配置型態能夠有效地提昇系統整體的可靠度。換言之，序列式系統架構及系統可靠度索引表能夠用以協助分析系統可靠度及建構具有高可靠度之主從式資料庫系統。

In the hardware system, N-tuple modular redundancy (NMR) has been widely used for raising hardware reliability. On the other hand N-version diversifying redundancy (NVDR), as a strategy for enhancing software reliability, has been proposed. Reliability is the focal point in the quest of flawless operations. Not only software but also hardware reliability affects the efficiency and exactitude of data manipulation in the client/server DBMS. There exists consensus for a trustworthy reliability model for specifying, evaluating, and raising the hardware reliability. In contract, NVDR remains a strategy for enhancing software reliability, and a specific and integrated representation that specifies a particular NVDR configuration is still needed. This thesis establishes the framework for representation of NVDR that facilitates reliability analysis.
Achievements of this thesis follow:
1. We characterize the physical configuration of client/server database system by addressing essential cause-effect factors that directly affect the reliability of DBMS operations.
2. We propose a representation of NVDR that offers a clear and definite framework for reliability characterization of a particular client/server NVDR database system.
3. We establish three reliability index tables, one for each of server-end, network, and client-end. These tables are used while developing the most reliable client/server database system architecture with resource constraints.

MANDARIN ABSTRACT i
ENGLISH ABSTRACT ii
MANDARIN ACKNOWLEDGEMENT iii
TABLE OF CONTENTS iv
LIST OF TABLES vi
LIST OF FIGURES vii
CHAPTER 1 INTRODUCTION 1
CHAPTER 2 A SIMPLE CLIENT/SERVER SYSTEM 3
CHAPTER 3 RELIABILITY ENHANCED DATABASE SYSTEM 13
3.1 Redundancy approach for software version 13
3.2 Redundancy for client/server DBMS structure 19
3.3 Physical framework 25
3.3.1 Physical framework for server-end 25
3.3.2 Physical framework for Network 26
3.3.3 Physical framework for client-end 26
3.3.4 Integrated physical framework 28
CHAPTER 4 REPRESENTATION FRAMEWORK 30
4.1 Model of series-framework 30
4.2 System configuration representation 37
CHAPTER 5 OPTIONS FOR RELIABLE CLIENT/SERVER DBMS ARCHITECTURE 43
5.1 Reliability of the server-end system 43
5.2 Reliability of the network 48
5.3 Reliability of the client-end system 49
CHAPTER 6 DEMONSTRATION EXAMPLES 51
6.1 Example I 51
6.2 Example II 59
CHAPTER 7 CONCLUSIONS AND DISCUSSIONS 63
REFERENCES 65
APPENDIX A: SYSTEM RELIABILITY INDEX TABLES 67
MANDARIN BIOGRAPHY 72

[1] Ashrafi, N., O. Berman, and M. Cutler, “Optimal design of large software-system using N-version programming,” IEEE Trans. Reliability, Vol. 43, No. 2, pp. 344-350, Jun. 1994,.
[2] Avizienis, A., and L. Chen, “On the implementation of N-version programming for software fault tolerance during execution,” Proc. IEEE COMPSAC 77, pp. 149-155, Nov. 1977.
[3] Berson, A., Client/Server Architecture. second edition, McGraw-Hill, Addison-Wesley Publishing Com., 1996.
[4] Brilliant, S.S., J.C. Knight, and N.G. Leveson, “Analysis of faults in an N-version software experiment,” IEEE Trans. Soft. Eng., Vol. 16, No. 2, pp. 238-247, 1990.
[5] Dahll, G., and J. Lahti, “An investigation into the methods of production and verification of highly reliable software,” Proc. SAFECOMP 79, 1979.
[6] Dewire, D.T., Client/Server Computing. McGraw-Hill, Inc., 1993.
[7] Ghribi, S., A. Abdennadher, and A. Jaoua, “Increasing software reliability using a signatrue method,” Information Sciences, Vol. 99, pp. 235-246, 1997.
[8] Gmeiner, L., U. Voges, “Software diversity in reactor protection systems: an experiment,” Safety of computer Control Systems, Pergamon, New York, 1980.
[9] Grubel, G., “The ANDECS CACE framework,” IEEE Control Systems Magazine, Vol. 15, pp. 8-13.
[10] Hatton, L., “N-version design versus one good version,” IEEE Software, Vol. 14, No. 6, pp. 71-76, Nov.-Dec. 1997.
[11] Heiler, S., and A. Rosenthal, “Engineering databases, tools, and management: an integration framework,” Proc. COMPCON 89, pp. 431 — 437, 1989.
[12] Park, H.C., W.B. Lee, and T.G. Kim, “A relational algebraic framework for models management,” Simulation Conference Proceedings, pp. 649 — 656, Dec. 1994.
[13] Chen, J.W., Implementation Considerations of Reliability-Enhanced Database (RED) by Diversifying Redundancy, National Cheng Kung University, Tainan, R.O.C., Jul. 1998.
[14] Verma, K.A., and T.M. Tamhankar, “Reliability-based optimal task-allocation in distributed-database management systems,” IEEE Trans. Reliability, Vol. 46, No. 4, Dec. 1997.
[15] Geiger, K., B. Tschumper, and J. Zander, Inside ODBC, Microsoft Press, 1995.
[16] Lyu, M.R., Y.T. He, “Improving the N-version programming process through the evolution of a design paradigm,” IEEE Trans. Reliability, Vol. 42, No 2, pp. 179-189, Jun. 1993.
[17] Ramez, E., B. Shamkant, Fundamentals of database systems, second edition, p. 707, 1994.
[18] Silberschatz, A., H.F. Korth, and S. Sudarshan, Database Systems Concepts, McGraw-Hill Companies, Ins., 1997.
[19] Vaughn, L.T., Client/Server System Design and Implementation, McGraw-Hill, 1994.