Info > Data Mining > 3-1. DM and DW > The current state
▷▶ The current state
☆ What is the current state of DW and how does this affect DM effort
?
☆ "data warehouse" refers to many different types of systems that have
little in common other than the fact that they contain data.
▶ Some common types of systems
▷ middleware tools
▷ departmental DW store
▷ interdepartmental staging area
▷ operation data store
▶ Middleware tools
▷ provide a single interface to a distributed network of source systems departmental DW or operational data stores.
▷ the disparate source systems do not to each other easily, not contain
compatible data, not designed for the complex queries needed for decision support.
▶ Departmental DW store
☆ summarized representations of interest to a single application area.
▷ implemented using OLAP, custom-built application.
▷ deliver value to the end user.
▷ do not solve the data integrity problem.
▶ Interdepartmental staging area
☆ cleanse data but maintain no history and do not have a logical data model.
☆ help solve the data cleansing problem.
▶ Operational data stores
▷ data from one or a set of related operation system → a relational database
▷ easier to cleanse and access the raw data for DW.
▷ not necessarily bring disparate data sources together in one place.
▶ Overgrown prototypes of large relational database contain the data but not in a format easily accessible.
▷ prototypes
DW의 개념을 보여주고 완전하게 구현됐을 때의 이점을 수치로 보여줄 수 있는 기반을 마련하기 위해 시험적으로 구축하는 소규모의 별도 DW. Data의 하부세트만 포함해 극히 제한된 비즈니스 문제만 해결한다. 프로토타입은 ‘뜨끈뜨끈한’ 살아있는 Data로 작업하며, 완전한 시스템에 필요한 사항들을 파악해 빨리 준비할 수 있도록 아주 신속하게 구축돼야 한다.
▶ The multitiered approach
※ a fairly normalized relational database with consistent logic model
that supports metadata, data marts, and end users.
☆ The last of these systems focuses on the data and how it used in the organization.
Info > Data Mining > 3-1. DM and DW > The current state