Abstract

Enterprise Planning is a structured approach to help a corporation establish an information system plan at a high enough level of abstraction to model the primary business sub-systems and applications. Whether the planning technique is the popular IBM Business Systems Planning (BSP) (IBM, 1984) or one of its derivatives, such as strategic information systems planning (Lederer and Sethi, 1992) the deliverables include the identification of the major business processes, their associated data classes, and the applications to which the business processes and data classes belong. The association among these three elements in BSP is referred to as the Information Architecture. The assumption is that the planning deliverables will be used throughout the rest of the SDLC, beginning in a top-down fashion in the analysis phase of development. The first stage of database development following enterprise planning is the identification and specification of subject databases and the development of a logical model to support the conceptual schema and its external schemas. The development of subject databases is frequently performed with a bottom-up approach, using hierarchical clustering in which entities are grouped together into databases according to their common participation in business processes. Tools such as enhanced ER diagrams (Elmasri, 1994) are most commonly used for the logical model. The development of the conceptual view and its implementation as the conceptual schema is also a bottom-up effort in that it is, at the least, the integration of all of the user views. It is tempting to carry out both of these database development efforts in a purely bottom-up approach, with little reference to previous enterprise planning. The temptation to disregard what has occurred during planning becomes greater with increased emphasis on prototyping and the misuse of rapid application development. The difficulty of applying data modeling to the entire enterprise has been noted (Scheer and Hars, 1992). The top-down/bottom-up dichotomy creates a coupling mismatch between enterprise planning and logical database development. This has been noted by many authors and with respect to BSP by Barlow, 1990. This mismatch can result in one of two reactions. The first is a reliance on ad hoc methods to couple planning and analysis. The second is to de-emphasize the outcome of the planning process or worse --to merely give lip service to it

Share

COinS