KIMBALL VERSUS INMON MODEL
Data warehousing refers to analysing and reporting the processed data in an organisation in order to enable the management to plan the organisational activities effectively. In order to increase the core competencies of the business, data warehousing tries to create an intelligent environment so that all the relevant data can be processed without any fuss. The reports are analytical in nature and help the management to track the operations in and out of the organisation efficiently. In this paper, the critical assessment of different data warehousing models is going to find a place in order to enable a regional university to build a strong data warehouse structure.
The Kimball model introduced the world to a new process of data warehousing technique called, dimensional modelling. Mishnaet al. (2016) have opined that this model is vastly different from the entity relationship model as, dimensional model does not have a wherewithal with relational database. Dimensional model got its inception because the data warehousing structure wanted to pay heed to the demands of the end users. That applies to the numerous number of queries from the end users that sometimes used to clog the entire system. The previous versions did not have the inventory to quench the queries of the end users. Dimension modelling sought to end this in an effective way. According to Kimball and Ross (2013), dimensional approach uses the facts and dimensions to design the entire structure. While, facts are most of the time numbers, dimensions are the cascade of information that flows through the database giving those numbers their desired value. The dimensional model has a four-pronged approach such as, choosing the business process, announcing the grain, finalising the dimensions and finding the right facts. Therefore, one of the most important aspect of using this model is that it is very easy to understand and has innovative structuring to handle complex operations easily. The delivery time is also quite less compared to other models.
On the other hand, Inmon model proposes a large warehouse structure for simplifying the entire data gathering and processing tasks. However, Elet al. (2015) have opined that this model being a large one, takes longer time than the previous one to deliver the outcome. In addition, from the point of view of constraining budget, Inmon model is more expensive. The approach taken in this model is more iterative one and therefore, the entire structure works upon the idea of top down approach. In this case, the regional university, being a large one, may find its problems solved in this model as, departmental data is processed more accurately here. However, the matter of complexity remains at large. Inmon model is more about data mart, whereas, Kimball is more about data warehouse. The basic difference between the two approaches is that data warehouse deals with multiple subjects. However, data mart deals with single core subject area. Inmon model is non-volatile in nature and all the changes made in this data structure are effectively monitored and put to a log. Therefore, from the aspect of privacy and security, Inmon model is more secure (Benny Austin, 2010).
As opined by Kimball and Ross (2013), facts and dimensions are beautifully inclined in this model so that the users can gain maximum amount of outcome from this model. In addition, the entire structure is dependent on bus architecture. Therefore, the data may not be categorized according to the attributes and relationships but it is far more simplified in nature. However, one of the most serious disadvantages of this model is that it does not adhere to the normalization theory. Therefore, Elet al. (2015) have opined that physical remodelling of this data structure is not possible. This model is more business-process oriented in nature and works with multilayer approach. However, this is a bottom-up approach and therefore, needs to flow well from the top level of an organisation. That applies to the fact that the end users have the liberty of forming queries to seek for a particular outcome. Ejimaet al. (2015) have opined that integration among various departments can be achieved only if, conformed dimensions are at place. The model is more like a star model and strides for achieving uniformity among the various connecting aspects by having equal correlation among them. The organisational requirement is tactical in nature and data integration model best works for departmental approach. In this case, a regional university is looking to bring all its departments under a single roof and for that reason, Kimball model looks to be suitable (1keydata.com, 2016).
Azizet al. (2015) have pointed out to the advantage of using Inmon model by stating that this is a part of bigger corporate information structure and takes all the existing database under its aegis. The everyday operations are atomic in nature. The parent organisation can track and record even the minor transactions so that transparency and privacy of the system reigns. The entire approach is subject oriented and for that reason, the end users may not have the chance to seek for response. Kimball and Ross(2013) have opined that ERD has the capability to redefine the relationship among various departmental attributes. Therefore, interdepartmental approach looks stunning. The structure of this model is primarily inclined to non-metric data and meets the requirement of multiple varied data structuring attributes. However, one of the most advantageous aspect of using Inmon model is that it provides the users the scope of changing and reforming the existing structure seamlessly.
Having gone through both the data warehouse structure, the paper has put forth all the advantages and disadvantages effectively to evaluate the best available option for the university to use for its data warehousing activities. As the discussion has revealed, both the approaches have their own advantages and disadvantages, However, keeping the aspect of the regional university and the relatively big size of itin perspective, it seems more logical that Kimball model will be suitable for integrating the five departments of the university seamlessly.
1keydata.com. (2016). Bill Inmon vs. Ralph Kimball. Available at: http://www.1keydata.com/datawarehousing/inmon-kimball.html [Accessed 1 Sep. 2016].
Aziz, A.A., Iliyasu, M.S., Mahmoud, I.A.K., Bongsu, R.H.R. and Jusoh, J.A., (2015). “A conceptual framework for a multidimensional model of talent management data warehouse”. Journal of Theoretical and Applied Information Technology, 80(2), p.334-356.
Benny Austin. (2010). Kimball and Inmon DW Models. Available at: https://bennyaustin.wordpress.com/2010/05/02/kimball-and-inmon-dw-models/ [Accessed 1 Sep. 2016].
Ejima, S., Kaneko, T., Ohta, Y. and Fehske, H., (2014). “Order, Criticality, and Excitations in the Extended Falicov-Kimball Model”. Physical review letters, 112(2), p.6401-7083.
El Moukhi, N., El Azami, I. and Mouloudi, A., (2015), “Data warehouse state of the art and future challenges. In Cloud Technologies and Applications (CloudTech)”, International Conference on (pp. 1-6). pp.23-61
Kimball, R. and Ross, M., (2013). The data warehouse toolkit: The definitive guide to dimensional modelling. Hoboken: John Wiley & Sons.
Mishna, F., Fantus, S. and McInroy, L.B., (2016). “Informal Use of Information and Communication Technology: Adjunct to Traditional Face-to-Face Social Work Practice”. Clinical Social Work Journal, pp.1-7.