| With the explosion of networks, computer professionals are being challenged with possibilities of new interoperation of systems, data, and software. This increased connectivity has inspired visionaries (BROD92a) (WEID93) to call for more intelligent cooperation among systems beyond simple sharing of software and data as currently afforded by language interoperability tools and heterogeneous database management systems. Richer interoperation requires ensuring relationships, semantics, and context surrounding an exported event are preserved before proper use by an interoperating component is possible. For example, many organizations have made significant investments in legendary systems but find themselves needing to extract relevant data and functionality for use with other legendary and new systems. Solutions when pursued, frequently are difficult to develop, complex to understand, and limited in their flexibility and ability to evolve with changing interoperation needs.; Our research contributions are threefold. First, we analyzed diverse platform needs to present a taxonomy of the interoperability problem space while focusing on the requirements for intelligent interoperation. Secondly, we defined a generalized approach to designing interoperation solutions to overcome shortcomings of current methodologies. We founded the concept of active modules which allow flexibility in defining interoperation events, conditions of interest, actions to take, and auxiliary support data. With this rule-based approach, we separated logic and reasoning found in individual rules from execution models that apply them. This permits designing tailored active modules for handling distinct interoperation tasks. Lastly, in this research we developed and experimented with a coherent family of interoperation mechanisms for supporting active modules which is based on active database technology and its rule-based paradigm. Several extensions to active database technology were needed to fully realize our approach. We describe our novel extensions for the grouping of rules, the control of rule execution, and the creation of associative stores. We also detail the conceptual and experimental analysis for three diverse applications requiring sophisticated cooperation. Results demonstrate the effectiveness of the employment of our methodology and prototype system as a general interoperability tool. |
