You are here


27 August, 2015 - 16:42

Relationships are associations between entities. As we have discussed in the previous section, a database consists of several (or many) different types of entities. However, in order to make the data stored in these entities effective for users to reconstruct descriptions of various business events, the various entities must be logically linked to represent the relationships that exist during such business events. The ease with which a user can ultimately extract related data from a database is heavily dependent on the quality of the database’s logical design—that is, effective identification of the relationships between different entities. These relationships map and define the way in which data can be extracted from the database in the future. The mapping of the relationships between entities (i.e., development of the E-R diagram) provides a roadmap for getting from one piece of data in the database to another related piece of data.

Review Question

Why is it important that you identify all of the important relationships when developing an entity-relationship (E-R) diagram?


A three-step strategy is generally most effective in identifying all of the relationships that should be included in a model. First, consider the existing and desired information requirements of users to determine if relationships can be established within the data model to fulfill these requirements. Second, evaluate each of the entities in pairs to determine if any entity provides an improvement in the describing of an attribute contained in the other entity. Third, evaluate each entity to determine if there would be any need for two occurrences of the same entity type to be linked—e.g., identify recursive relationships. Appendix A describes the development of an ER model in greater detail.


A major thrust in many organizations has been a move toward completely integrating all data across an organization. These completely integrated enterprise models are the foundations for implementing enterprise systems, which are discussed in Technology Insight 3.2. Integration allows many users to share entity-level data by linking business events within related business processes.


Technology Insight 3.2

Enterprise Systems

Enterprise systems are integrated software packages designed to provide complete integration of an organization’s business information processing systems and all related data. These systems are based on event-driven systems concepts, which include the capturing of business data for supporting decision making, as well as integration of the underlying data to facilitate access and ad hoc analysis.

A number of enterprise systems are commercially available. The dominant player is System Application Products (SAP) R/3, which commands the largest percentage of the Fortune 500 market. Several other products are available and have established large customer bases—often through establishing excellence in certain market niches. These other vendors include JD Edwards, PeopleSoft, and Oracle. While these products are designed to offer integration of everything from accounting and human resources to manufacturing and sales staff logistics, products designed to focus on specific industries are also appearing in the marketplace. These systems are capable of extracting data from both enterprise systems’ data sources and legacy systems that may still exist within an organization (or subsidiary of the organization). They can also support a Web interface to allow business partners to initiate business events directly.

Originally, the implementation of enterprise systems was predominantly targeted at large multinational manufacturers such as General Motors, IBM, and General Mills. This strategy aimed where benefits would be expected to be the greatest, in that large multilocation and multidivision companies often present the greatest challenges to managers who mine data from corporate databases to improve overall organizational decision making. Enterprise systems allow companies to standardize systems across multiple locations and multiple divisions in order to link data in a consistent fashion and provide organization-wide accessibility.

Large enterprises were the predominant implementers of enterprise systems, but largely due to the costs of implementation. These systems typically took a year or more to implement at a cost of up to hundreds of millions of dollars, necessitating a similarly significant return in benefits. As advances in technology underlying these systems has evolved, small and medium sized enterprises (SMEs) have driven the new implementation base. This shift has happened primarily due to two drivers: (1) the move towards web-browser driven systems that reduce the expense of both the technology and training; and (2) the emergence of application service providers (ASPs) that implement enterprise systems and then lease out use of the enterprise system to several other companies. In other words, the ASP runs the hardware and software for the company that wants its data integrated via an enterprise system, and the company saves money by essentially sharing the costs of the enterprise system implementation and maintenance with several other companies that also use the same ASP. ASPs are discussed in greater detail in Chapter 7 of the text.