• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10c
• /
• pp.262-264
• /
• 2002

데이터웨어하우스는 대용량의 데이터를 대상으로 복잡한 분석을 수행하거나 전략적 의사 결정을 위해 사용되기 때문에 신속한 질의 응답을 위한 성능 관리가 무엇보다 중요하다. 데이터웨어하우스는 일반 운영계 시스템과는 다른 목적과 특성을 가지기 때문에 그에 적합한 성능 모니터링 방법이 필요하다. 본 연구에서는 오라클 데이터베이스를 위한 OLTP(On-Line Transaction Processing)용 국산 성능 모니터링 도구인 DBMax를 데이터웨어하우스 환경에 적용할 수 있도록 기능을 확장한다. 우선 기존의 운영계 시스템과는 구분되는 데이터웨어하우스의 특성을 토대로 오라클9i에서 제공하고 있는 성능 항상 기법인 실체화된 뷰(MV: Materialized View)와 EH(Data Extraction, Transformation and Loading) 화장 기능을 모니터링하기 위한 요구사항을 분석하고 이를 지원하기 위한 DBMax 확장 방안을 설계 및 구현한다. 실체화된 뷰와 ETL 확장 기능을 위해 제공되는 오라클9i의 다양한 스키마 객체에 대한 정보 및 성능 관련 지표를 제시하여 SQL 튜닝 기능을 지원하며 사후 분석을 위한 DBMax의 로그 파일에서 의미 있는 SQL문을 추출하여 잠재적으로 유용한 실체화된 뷰를 권고할 수 있도록 기능을 확장한다.

• The KIPS Transactions:PartD
• /
• v.10D no.3
• /
• pp.407-416
• /
• 2003

As the usage of database systems dramatically increases and the amount of data pouring into them is massive, the performance administration techniques for using database systems effectively are getting more important. Especially in data warehouses, the performance management is much more significant mainly because of large volume of data and complex queries. The objectives and characteristics of data warehouses are different from those of other operational systems so adequate techniques for performance monitoring and tuning are needed. In this paper we extend functionalities of the DBMax, a performance administration tool for Oracle database systems, to apply it to data warehouse systems. First we analyze requirements based on summary management and ETL functions they are supported for data warehouse performance improvement in Oracle 9i. Then, we design architecture for extending DBMax functionalities and implement it. In specifics, we support SQL tuning by providing details of schema objects for summary management and ETL processes and statistics information. Also we provide new function that advises useful materialized views on workload extracted from DBMax log files and analyze usage of existing materialized views.

• Journal of Korean Society of Rural Planning
• /
• v.11 no.4 s.29
• /
• pp.75-91
• /
• 2005

This study designs an integrated data architecture to systematically manage the agricultural statistics database. Managing the agricultural statistics is important since it provides data for policies and decision making for agribusinesses. Ministry of Agriculture and the National Statistical Office collect the basic agricultural statistic data which provides the basis of logical decision making and agricultural policies. However, the agricultural statistic data has not well been used. The data has not been consistently collected nor managed. The raw data has not been organized nor processed to meet various demands. The needs has been arisen for a consistent agricultural statistics system to increase the relevance, accessibility, and efficiency of data for various users. There are massive amount of data accumulated over a long time period. Introducing the new system and reorganizing the data will bear large risks. A systematic method is required to reduce the risks in planing, building, and maintaining the database without hindering administration. This study provides a design of the agricultural statistics system architecture based on the user requirement analysis (URA) and similar systems abroad. We have also build a prototype to check the implementability of the system design.