Bus architectures are based on the concept of conformed dimensions in order to enable the creation of data warehouses in an iterative manner, and to ensure the consistency of data between the different business processes analyzed.

Context

In computing, a bus is a mechanism enabling the transmission/sharing of information between several components.

Dimensional modeling is a strategic planning activity in the sense that it is used to model different stakeholders’ needs while maintaining the consistency of data between the different datamarts.

To this end, each datamart, designed on the basis of a star schema, shares some of these conformed dimensions with other datamarts (see the diagram below).

Source du schéma : Kimball, Ralph, and Ross, Margy. The Data Warehouse Toolkit: The Complete Guide to Dimensional Modeling.

In this diagram, we can see that the business processes for Store Sales and Store Inventory management share three conformed dimensions: Date, Product and Store. This ensures the consistency of analyses comparing inventory and sales, for example.

Bus matrix

Source du schéma : Agile Data Warehouse Design: Collaborative Dimensional Modeling. Laurence Corr and Jim Stagnitto.

The bus matrix provides a global overview of all the data in the decision-support information system. This essential document helps to establish a clear dialog on data with your interlocutors.

Each row of the matrix represents an organizational process that is associated with a datamart. In fact, the datamart is simply a fact table surrounded by its dimensions.

That is why each of the columns in the bus matrix represents the different dimensions present in the system. When a dimension is used by several fact tables, it is said to be conformed.

Reading Bus matrix: Star Schema. Christopher Adamson. pages 100-101

Reading Bus matrix: Agile Data Warehouse Design: Collaborative Dimensional Modeling, from Whiteboard to Star Schema Laurence Corr, and Jim Stagnitto.pages 101-102, 128

Reading Event matrix: Agile Data Warehouse Design: Collaborative Dimensional Modeling, from Whiteboard to Star Schema Laurence Corr, and Jim Stagnitto.pages 102-103

Bus architecture

Source du schéma : Agile Data Warehouse Design: Collaborative Dimensional Modeling. Laurence Corr and Jim Stagnitto.

This diagram shows the data transformation flows in a bus architecture. Each event source undergoes an extraction and transformation in the ETL zone (to which users are denied access as the consistency of the data has not yet been ensured), before finally being loaded into the data warehouse where the data is historized, consolidated and its consistency is ensured, notably by means of conformed dimensions.

Here we can see that a datamart is not reserved for a single department of the organization but meets a global need that can be shared and interpreted in the same manner by different stakeholders in the organization.

Reading Bus architecture: Star Schema. Christopher Adamson. pages 100-104

Reading Bus architecture: Agile Data Warehouse Design: Collaborative Dimensional Modeling, from Whiteboard to Star Schema Laurence Corr, and Jim Stagnitto.pages 100-103, 128