Matching client/server processing architectures with information processing requirements

The 1990's are witnessing the rapid gyraterowth of Client/Server (C/S) computing. This growth can be attributed to the many benefits which C/S systems provide to organizations. These benefits include, among others, the potential to improve business practices, cost savings, and increased productivity. Achieving these benefits, however is not an easy task, since there are many types of C/S processing architectures which system designers can select. These processing architectures are derived from the method by which the C/S processing resources (presentation functions, business logic, and data management) is split between the client and server. Therefore for an organization to benefit from the C/S model, it should ensure that its processing architecture matches its information needs. Researchers have pointed out that organizations in transition to the C/S model should identify their information requirements, and then determine the best C/S architecture that supports them. This study examines the appropriate match which is required between a workgroup's information processing requirements and C/S processing architectures to create an effective system. The theoretical basis of this study stems from Galbraith's (1977) information processing theory, which states that an systems performance is contingent on an appropriate match between context variables such as task characteristics, and structural designs such as information system design. The independent variables used in this study are (a) task characteristics, which have been identified by researchers as one of the primary determinants of information processing requirements; and (b) C/S processing architecture, which determine the information capabilities of an organization. The dependent variable is C/S effectiveness. The sample data required for this study obtained from C/S managers and C/S users in variety of industries, through a combination of archival data, telephone interviews, and a mailed survey. The unit of analysis in this study is the workgroup. Two hundred and ten (210) sets of questionnaires were mailed to eighty seven (87) usable datasets were returned. The data was analyzed using moderated regression analysis. The matching concept was measured through the interaction terms of the task characteristics, namely, task structurability and task interdependence, and C/S processing architecture. The interaction terms were statistically significant, suggesting that selecting an appropriate C'S processing architecture that matches the information processing requirements is important for the effective performance of a C/S system. In addition this study provides guidelines for systems developers in designing C/S systems.