Data Warehousing Special Topic (22): Bus architecture and dimensional modeling benefits-Miscellaneous

First, the bus architecture

in the Data Warehouse of dimension modeling, there is a concept called Bus Architecture, the Chinese is generally translated as "Fieldbus architecture". The bus architecture is one of three key concepts in the Multidimensional architecture (MD) of Kimball, and theother two are the consistency dimension (conformed Dimension) and the consistency fact ( Conformed Fact).

in a multidimensional architecture ( MD) in the Data Warehouse architecture, the dominant idea is to step up the Data Warehouse, which is composed of data marts into the enterprise Data Warehouse. However, before building the First Data mart, the first thing architects need to do is to design standardized dimensions and facts that are uniformly interpreted throughout the enterprise, namely, conformance dimensions and consistency facts. And the development team must strictly follow this architecture for the iterative development of the data mart.

A consistent dimension is like a group of buses within an enterprise, and the fact of a different data mart is like a component plugged into that set of buses. This is also the reason for calling the bus architecture.

in the actual design process, we usually put the bus schema list into a matrix form, which is classified as a consistent dimension, the behavior of different business processes, that is, the fact that the intersection points marked that the business process is related to the dimension. This matrix is also called the bus matrix.

the bus architecture and the consistency dimension, the consistency fact, together make up Based on the multidimensional architecture of Kimball, a set of methodologies for building data warehouses can be established. Because the bus architecture is the core of the multidimensional architecture, we sometimes refer to the multidimensional architecture as the bus architecture.

second, the bus matrixTypically, a row of a bus matrix produces several related fact tables, which enable you to track business processes from different angles. The order business process may have a line item level order Transaction Fact table and order-level order snapshot fact table. These two order-based dimension models belong to the order business process, which is called the business Process dimension model.

three, the advantages of dimensional modeling

The Data Warehouse uses the benefits of using dimensional modeling: Easy to understand, high performance for queries, flexibility and scalability for modifications.

Dimension modeling is a process that can be continuously augmented and added

(1) Add dimensions to the existing fact table.

(2) Add facts to the fact sheet.

(3) Add attributes to the dimension table.

In the comparative understanding of the business situation, you can start with the bottom-level fine-grained construction, and vice versa, start with the coarse granularity of business demand, to top down;

Data Warehousing Special Topic (22): Bus architecture and dimensional modeling benefits-Miscellaneous