Data Warehouse Models Data models enterprise data models data warehouse models industry data models generic data models best practice data models Erwin Allfusion data models HP Oracle EDS

The data warehouse is the central repository for corporation information representing the integrated data requirements of the enterprise and designed to support the analysis, DSS and reporting requirements of the entire organization. It has characteristics unique to its function:

>  Based upon a comprehensive enterprise model
>  Provides integrated data to the organization
>  Serves a broad user community
>  Contains cleaned, consistent data
>  Data is at granular level of detail
>  Significant or often-used data is summarized
>  Contains time-based "historical" data
>  Driven by analytic requirements
>  Structured by aligned business areas
>  Addresses evolving information needs
>  Is not updated
>  Contains summary data

The challenges in designing and building the data warehouse lie in the details related to what the components of the data warehouse will be and structuring them consistent with business practice.

Verbally, we know the things that need to be the components of the data warehouse - "GL accounts, customers, orders, channels, products, BBB..." and as many related pieces of data as can be conceived by business function. The challenge is to now identify the best-practice data structures consistent with business practice and supporting legacy information requirements. What is needed is a 'blueprint' or picture of the target data structures that can be used to load current data, integrate legacy data and establish a data architecture supporting the needs of the organization into the future.

Any failure of the data warehouse normally lies in its design being driven by the data warehouse effort and not the business requirements themself.

In order to identify the needs of the business at both a high-level and low-level of data granularity, comprehensive detailed data models are required. A comprehensive enterprise model paints the big picture with successive Business Area Models providing ever more detailed and comprehensive data representations.

This is the ADRM approach to building the data warehouse.

1. Start with a best-practice industry set of data models

2. Develop a corporate enterprise model taking advantage of the ADRM 80% fit to take advantage of both design and industry content. This model will be used for planning and to identify the iterative business data contents of the data warehouse.

3. Use ADRM Business Area Models as the source to validate data warehouse content, develop target data structures and identify points of data integration.

4. Apply the ADRM Data Warehouse Model as a jump-start to propagate the results of the above into the data warehouse model.

The ability to apply large amounts of proven content to an initial design that is graphically easy-to-understand and based upon CASE technology will slash development time, reduce costs and provide an immediate 'picture of the data warehouse'.

The 'blueprint' of the data warehouse at the earliest possible time is the single greatest contributing factor to the success of the data warehouse.

"Using Template Data Models
To Build The Data Warehouse"