Loading...
|
Please use this identifier to cite or link to this item:
https://nccur.lib.nccu.edu.tw/handle/140.119/64418
|
| Title: | UML導向之主從式物件資料庫績效評估之工作量模型 |
| Other Titles: | An UML-driven C/S OODBMS Workload Model for Benchmark |
| Authors: | 諶家蘭;方相宜
Seng, Jia-Lang;Fong, Ian-I |
| Contributors: | 會計系 |
| Keywords: | 物件技術;物件資料庫;主從式架構;績效評估;工作量模型
Object-oriented technology;Object-oriented database management systems;Client/Server structure;Performance evaluation and assessment;Benchmark;Workload model;UML |
| Date: | 2003-01 |
| Issue Date: | 2014-03-04 15:25:46 (UTC+8) |
| Abstract: | 物件資料庫技術自九十年代帶領資訊科技和網路科技進入整合和連結世代，隨著二十一世紀來臨，資料庫科技不斷演進，接受新世紀挑戰和洗禮。自結構化資料庫到平結構超文件資料庫到少結構化多資訊資源異質存取資料庫，再由資料倉儲、資料發掘到多媒體網路資料庫、XML文件資料庫和元件資料庫。我們都可以看到物件資料庫的應用和重要性，他們使得網路平台和交易平台能夠以資料獨立、流程獨立和平台獨立方式和分散式運算來達成。新興科技的成敗在技術面、管理面和市場面面面重要，如何提出比較客觀、公允、一致、正確、實用、量化的評估機制是非常重要的。績效評估自1960年代開始有系統性被開發，學界和業界投入大量心力研發改進，使得績效評估成為最重要最有力的評估機制。資料庫領域中，績效評估一直扮演關鍵決定因素，而其中最有價值的是工作量模型的研究和開發。在物件領域中，主流和標準績效評估方法一直以問題領域設定和問題領域專屬為工作量模型的基礎，所產生的績效數據，一直為使用者、管理者、開發者的重要參考。但是，長久以來，使用者需要重新客製，才能真正使用標準數據。本研究期望針對研究瓶頸，提出以使用者工作量需求為導向，使用標準載具，透過物件需求分析方法，承載資料需求、交易需求、控制需求，將不同使用者需求，以系統化和模式化，先以圖形型態UML方法，再轉換成為文字型態語法語言，透過產生器和驅動器自動產生使用者測試套組。
Object-oriented database technology enables the integration of information technology and Internet technology. It facilitates the new and emerging development of data warehousing, data mining, multimedia network systems, XML document technology, multi-source Web information access, interoperability of heterogeneous Web databases, and mobile databases. Performance evaluation and assessment plays the vital role of the success of these new and emerging technologies. It is a scientific and quantitative method to analyze and assess the strong points and weak points of technology since 1960s. In the database area, benchmark is the name of performance study. Standard and major benchmark methods are domain specific and domain dependent. Users, managers, and developers use these benchmark results as a reference to make decisions on troubleshooting, procurement, breakthrough, and bottleneck detection. However, a re-occurring problem of using these benchmark results is that these statistics are only estimates. How close these numbers are to the actual performance depends on how close the standard workload is to the actual applications. When your problem domain differs from the standard domain, performance results vary and become misleading and meaningless. Workload is the core issue of the re-occurring problem. In this research, we perceive workload model as a systematic modeling of user requirements in applications. The process of workload requirements analysis and modeling becomes the core of a benchmark study. The workload model can be more open and more generalized. We adopt UML as the diagramming analysis technique. We develop corresponding textual and grammatical representation. Prom analysis, specification, to generation and execution, we present a coherent and cohesive requirements-driven benchmark model that is more representative, correct, and real. |
| Relation: | 資訊管理學報, 9(2), 1-30 |
| Data Type: | article |
| Appears in Collections: | [會計學系] 期刊論文
|
Files in This Item:
| File |
Description |
Size | Format | |
|---|
| 130.pdf | | 1813Kb | Adobe PDF2 | 1146 | View/Open |
|
All items in 政大典藏 are protected by copyright, with all rights reserved.
|
