Vous êtes sur la page 1sur 2

Database Systems: Top-Down or Bottom-Up Design?

By: Yasmeen Shabana

BSc of Electrical Engineering and Power Machines.

MSc of Software Engineering.


_____________________________________________________________________________________________

Designing table structures is achieved through various designing strategies. The designer could
start by identifying a regular entity then break this into all the required entity subtypes based on their
distinguishing characteristics. Or could start by identifying multiple entity types then find common
characteristics of those entities that are suitable for creating a higher-level supertype entity. These two
approaches are called specialisation (top-down) and generalisation (bottom-up) (Coronel and Morris,
2015).

The top-down approach emphasis on the high-level identification of entity types, attributes, and
relations. It starts by examining the business rules, identifying the database objects within the domain,
and developing a conceptual model. This approach utilises relational or object-relational database
models (Kung, Kung and Gardiner, 2012). When developing a database for a company the designer will
start by modelling the company and its strategy, target customers, supply chain, marketing, policies,
competitors, and overall economy (Wirsch, 2014). This approach has a great advantage when designing
database from scratch as it gives visibility of the effect of further changes (Gadicha et al., 2012).
However, it is more time consuming when collect the data and requires more communication towards
the end user to fulfil all the requirements during implementation phase while focusing on a high-level
conceptual model in the designing phase (Elmasri and Navathe, 2011).

The bottom-up approach focuses on analysing the lower-level elements of the data such as
attributes and functional dependencies then works towards a logical model of the database by bundling
and categorising relevant attributes. It starts with collecting user’s requirements for data presentation
such as reports, specific patterns, and views. Normalisation is a commonly used methodology for
bottom-up strategy (Kung, Kung and Gardiner, 2012). Its purpose is to identify similarities among
different entities such as common attributes or relations (Connolly and Begg, 2004). When developing a
database system for a library, the designer would start by modelling the library based on existing
authors and their books then group the books by subject. This approach benefits from the normalisation
process to minimise data redundancy and produce efficient tables structures (Gould, 2015). It is a
powerful modelling tool that factors out shared specifications and limits design errors (Eder and
Kalinichenko, 2012). However, the high-level view of the database is not completely visible, therefore,
the effect of future changes is not predictable. As a starting point, it requires collecting all the binary
relationships among attributes which is almost impossible to do for all relationships of attributes pairs in
the database (Elmasri and Navathe, 2011).

Combination the two approaches is the best strategy for the database design. Using the top-
down strategy to outline the requirement and partition it into logical divisions where the schema part
for each partition is designed according to the bottom-up strategy. This mixed strategy gives great
flexibility and efficiency to develop a database model that is not only focusing on the tables, entities,
relations but first and foremost adhere to the problem domain and reflects the business model of the
organisation as a real-world representation.

REFERENCES

Connolly, T. M. and Begg, C. E. (2004) Database Solutions: A Step-By-Step Guide To Building Databases.
2nd edn. Harlow, England ; New York: Pearson/Addison Wesley.

Coronel, C. and Morris, S. (2015) Database Systems Design, Implementation, and Management. 11th
edn. Cengage Learning.

Eder, J. and Kalinichenko, L. A. (2012) Advances in Databases and Information Systems: Proceedings of
the Second International Workshop on Advances in Databases and Information Systems (ADBIS’95),
Moscow, 27–30 June 1995. Springer Science & Business Media.

Elmasri, R. and Navathe, S. (2011) Fundamentals of Database Systems. 6th ed. Boston: Addison-Wesley.

Gadicha, A. B. et al. (2012) ‘Top-Down Approach Process Built On Conceptual Design To Physical Design
Using LIS, GCS Schema’, International Journal of Engineering Sciences, 3(1), p. 7.

Gould, H. (2015) Database Design and Implementation: A Practical Introduction Using Oracle SQL. 1st
edn.

Kung, H.-J., Kung, L. and Gardiner, A. (2012) ‘Comparing Top-down with Bottom-up Approaches:
Teaching Data Modeling’, Proceedings of the Information Systems Educators Conference, New Orleans
Louisiana, USA, 29, p. 11.

Wirsch, A. (2014) ‘Analysis of a Top-Down Bottom-Up Data Analysis Framework and Software
Architecture Design’, p. 72.

Vous aimerez peut-être aussi