Kimball’s and Inmon’s Data warehousing approaches

Today data has become a very hot topic. Data is generating at high velocity and organizations are more focused on analyzing data to have business decisions. Ralph Kimball and Bill Inmon are two pioneers in data warehouse designs. They have suggested two philosophies to design a data warehouse.

Ralph Kimball’s approach is called the bottom-up approach. Also, it is called Dimensional modeling.

In his approach, we divided the business process into different parts and design data marts for each of these areas. These data marts are in 3NF. Also, this approach called bottom-up, because we are designing the final data warehouse using data mart design for a small section. There is an ETL process to design these data marts from the OLTP data sources. After designing the data marts, the second phrase is designing the data warehouse. The data warehouse is normalized and it includes dimensions and fact tables. The idea behind this approach is to use simple and fast queries to retrieve data. Then the data can be used for analytics and report generations.

Bill Inmon’s approach is called top-down. That means we design the full data warehouse first and then design data marts according to business analytics and report generating requirements.

In the designing process of this architecture, we need to consider the overall business process and need to have a good understanding of overall process. For example, When we designing a logical model it is needed to design a logical model this all the details to the customer and needs to include all the relationships and dependencies. Here is a key point to minimizing data redundancy. Therefore the data warehouse in 3NF. After reducing the complexity of quarries, Inmon has suggested data marts for each individual department. These data marts can be denormalized and using them can be easily generated reports.

In Inmon’s method, is time-consuming and needs a team of specialists with comparing Kimball’s method. Therefore the initial cost is high in Inmon’s methods and Kimball’s method won’t need a team of specialist. So the initial cost is low. But the maintenance of Kimball’s method is difficult and costlier than Inmon’s method. Inmon has suggested doing the data integration enterprise-wide and the Kimbal has suggested doing it for individual business areas.

By considering all the above-mentioned facts, It can be concluded that for small businesses and organizations, the suitable data warehouse designing method is Kimball's method. Because for those companies a data warehouse is not the main focus and they can’t afford that more and a budget to develop a data warehouse. But for big companies, the more suitable method is Inmon’s approach.