Developing the Atomic Warehouse Model

When you develop the Atomic Warehouse Model (AWM) in the context of a project, AWM is propagated from the enterprise level to the project level and finally to the practitioner level. Then, it is customized based on the scoped input elements, and the changes are harvested at the project level. When all practitioners have harvested their changes at the project level, the project AWM is harvested at the enterprise level. This results in a new version of your enterprise AWM.

Before you begin

This task assumes that the business terms are properly defined in the context of your project. You can have used your own format or the BFMDW proposed format in Information Governance Catalog (IGC). When the business terms are captured by using the BFMDW vocabulary in Information Governance Catalog (IGC), the mapping to the corresponding AWM elements is already done for any preexisting business term and this accelerates your development.

About this task

Multiple instances of this task can run in parallel in the following cases:
  • Multiple practitioners who are working on the same project.
  • Multiple projects that are running in the enterprise.
The frequency of propagation and harvesting activities is driven by the enterprise model management policies. However, frequently harvest changes to minimize the number of conflicts and accelerate the harvesting.

When developing AWM, you propagate its enterprise version into your practitioner version, you customize it, and once it has been consolidated at project level, you accelerate your new version back to the Enterprise stream.

Figure 1. Developing AWM tasks
Developing AWM tasks

Procedure

  1. Propagate the enterprise version of AWM into your practitioner AWM.
    See Propagating data models using AWM.
  2. Customize your practitioner AWM in your data modeling tool.
    See Customizing the Atomic Warehouse Model.
  3. Harvest the practitioner AWM at the project level.
    For each practitioner at a time, follow the steps in Harvesting data models.
  4. Harvest the projects AWM at the enterprise level.
    For each project at a time, see Harvesting data models.
  5. Import the enterprise AWM into InfoSphere Metadata Asset Manager (IMAM).
    See Importing the logical IDA data models into IMAM.
  6. Export the glossary assignments from IDA into IMAM.
    See Exporting the glossary assignments from IDA into Information Server.