The database components of a data warehouse can be described using either relational database modeling or dimensional modeling. A conceptual database model is a collection of logical constructs used to represent the database's data structure as well as the data relationship(s) found within that structure, (Rob & Coronel, 2004).

There are many different types of database conceptual modeling, which includes: entity relationship modeling (ERM), object oriented modeling (OOM), and semantic data models, such as object oriented data modeling (OODM), and extended relational data modeling (ERDM). The relational database model, which represents a global view of the database, is mainly and commonly illustrated using entity relationship (ER) model with the help of entity relationship diagrams (ERD). And internally, the designers or modelers match the conceptual model’s characteristics and constraints with that of the selected implementation model (Rob & Coronel, 2004).

The ERM complements the relational database model concepts by ensuring proper design of relational databases in a tightly structured design environment, thus providing easy to understand ad hoc query and set-oriented access capabilities and incorporating most semantics in data model. The ERM uses the ERDs to help identify the database's main entities, attributes, and their relationships, and the number of entity occurrences associated with an occurrence of a related entity (cardinality notations), (Kimball & Ross, 2002).

ER diagrams enable designers to easily map the ERM to the relational database model’s tables and attributes in a series of well-defined steps to generate all the required database structures; thus freeing them from the complexities associated with physical data representation. Likewise, because of the visual presentation of data and their relationships, end users find it easier to understand the ERM components and their roles (Rob & Coronel, 2004).

However, due to the real-world complexities of data and information presentation, the ERM sometimes does not accurately capture the complex data types and their relationships, which leads to loss of information content. Despite some disadvantages, ERM is still a valuable tool for relational data modeling. But dimensional modeling has become the favorite for many data warehouse and data mart database modeling at present.

Therefore, a good conceptual database data model should have the following characteristics: simplicity, data reusability, enforcement of business rules, non-redundancy, accuracy and completeness, stability and flexibility, and communicate effectively (Kimball & Ross, 2002).

References:

1. Rob, P. & Coronel, C. (2004). Database Systems: Design, Implementation, & Management. 6th Edition. Thompson Course technology

2. Kimball, R. & Ross, M. (2002). The data warehouse toolkit: The complete guide to Dimensional Modeling. 2nd Edition, illustrated. Published by John Wiley and Sons