Client-Server Computing, Cooperative Processing, Distributed Computing and other terms are used to describe approaches to information management where collections of typically heterogeneous computing hardware, referred to as the information architecture, are deployed to satisfy organizational information requirements. A common theme in these approaches is the interconnection of discrete computing resources to create a cooperative computing (CC) environment.; The growing popularity of these approaches mandates a re-evaluation of traditional criteria regarding the selection of an appropriate information architecture. Recent literature and research in CC system development has emphasized software related issues such as application partitioning, inter-process communications capabilities, graphic user interface development and the like. The suitability of the information architecture has been given inadequate treatment. Regardless of the development methodologies and deployment strategies used for information system implementation, organizations must still address a common set of fundamental issues. These issues are: (1) determination of feasible and appropriate configurations of computing capacity and data communications that should be allocated to organizational units to satisfy the overall information management requirements of the organization. (2) derivation of an information architecture to provide the best fit with respect to the organization's overall needs, in consideration of the potential constraints of limited resources and/or conflicting strategic goals.; Individual unit costs for computing machinery and data communications continue to exhibit dramatic declines over relatively short time periods. However, the aggregate and long-term consequences of information management capacity planning must extend beyond the initial cost of a particular component or architecture. A substantial body of research, along with a well defined set of market products, addresses planning and selection issues for discrete systems and centralized processing environments. Evaluation of complete CC environments remains a largely unexplored research domain.; This dissertation presents a knowledge-based approach for designing information architectures in heterogeneous, cooperative computing environments, and defines an open and extensible approach to deriving CC architectures. The approach is based on the technique of heuristic classification. The methodology developed in this research transforms formal specifications of an organization's information requirements into a series of technology-independent generic architectures. These represent the means of satisfying those requirements in the contexts of resource allocation and system design strategies typical in the cooperative computing domain. The generic architectures are subsequently translated into sets of contemporary technology components required to implement them. These sets of technology components may then be evaluated according to organizational preferences to determine the appropriate information architecture.; A prototype for a knowledge-based decision support system (DSS), the Information Architect, has been developed to serve as proof-of-concept for the approach. The approach and the prototype were tested in two organizations. Results indicate that the approach helped the organizations to specify appropriate architectures, and offer validation of the techniques presented in this research.
展开▼