To those who are unfamiliar with Ralph Kimball and Bill Inmon data warehouse architectures please read the following articles:
Kimball vs Inmon in data warehouse architecture
Both Kimball and Inmon’s architectures share a same common feature that each has a single integrated repository of atomic data. In Inmon’s architecture, it is called enterprise data warehouse. And in Kimball’s architecture, it is known as dimensional data warehouse. Both architectures have an enterprise focus that supports information analysis across the organization. This approach enables to address the business requirements not only within a subject area but also across subject areas.
However there are some differences in the data warehouse architectures of both experts:
- Kimball uses dimensional model such as star schemas or snowflakes to organize the data indimensional data warehouse while Inmon uses ER model in enterprise data warehouse. Inmon only uses dimensional model for data marts only while Kimball uses it for all data
- Inmon uses data marts as physical separation from enterprise data warehouse and they are built for departmental uses. While in Kimball’s architecture, it is unnecessary to separate the data marts from the dimensional data warehouse.
- In dimensional data warehouse of Kimball, analytic systems can access data directly. While in Inmon’s architecture, analytic systems can only access data in enterprise data warehouse via data marts.
Kimball vs. Inmon in data warehouse building approach
Bill Inmon
Bill Inmon recommends to build data warehouse that follows top down approach. In Inmon’s philosophy, it is starting with building a big centralized enterprise data warehouse where all available data from transaction systems are consolidated into a subject-oriented, integrated, time-variant and non-volatile collection of data that supports decision making. then data marts are built for analytic needs of departments.
Ralph Kimball
Contrast to Bill Inmon approach, Ralph Kimball recommends to build data warehouse that follows bottom up approach. In Kimball’s philosophy, it is first start with mission critical data marts that serve analytic needs of departments. Then it is integrating these data marts for data consistency through a so called information bus. Kimball makes uses of dimensional model to address the needs of departments in various areas within enterprise.
How to choose between Kimball vs Inmon approach for building data warehouse?
Here are the most important criteria how to choose between Kimball vs Inmon approach.
| Characteristics | Favours Kimball | Favours Inmon |
|---|---|---|
| Business decision support requirements | Tactical | Strategic |
| Data integration requirements | Individual business requirements | Enterprise-wide integration |
| Structure of data | KPI, business performance measures, scorecards… | Data that meet multiple and varied information needs and non-metric data |
| Persistency of data in source systems | Source systems arequite stable | Source systems have high rate of change |
| Skill sets | Small team of generalists | Bigger team of specialists |
| Time constraint | Urgent needs for the first data warehouse | Longer time is allowed to meet business’ needs. |
| Cost to build | Low start-up cost | High start-up costs |
In this article, we’ve discussed about the Kimball vs Inmon in data warehouse architecture and design approach. In addition, we’ve provided the information that you can choose between Kimball vs Inmon to build your data warehouse.
