A Metastore-based Data Warehouse Development Methodology

Abstract

Data warehouse (DW) is important for analytical processing. Metadata is a key to its architecture. This paper proposes an architecture that consists of seven components. To illustrate data warehouse environment(DWE), this paper proposes taxonomies having four flows. On the basis of the taxonomies and metadata, this paper proposes a methodology for building the data warehouse and metadata simultaneously. This integrated development methodology (IDM) consists of seven phases: (ⅰ) preparatory phase, (ⅱ) requirement analysis phase, (ⅲ) data warehouse development phase, (ⅳ) operational data store development phase, (ⅴ) data mart development phase, (ⅵ) metastore development phase, and (ⅶ) maintenance phase. A metastore system is proposed to help develop metadata interactively. An illustrative example is investigated to demonstrate the usefulness of IDM.