An intelligent mediator-based framework for enterprise application integration

Enterprise Application Integration (EAI) is a major problem confronting manufacturing organizations that have deployed large-scale enterprise information systems and aim to coordinate inter- and intraorganizational product development, supply-chain, and customer management activities. Successful EAI is a key enabling step toward implementation of viable Product Life-cycle Management (PLM) strategies. Extant EAI technologies, such as distributed object and messaging technologies for communication and ontology-based database schema integration, are tedious to develop and maintain in an organization. Though workflow management systems have enabled process coordination of both manual and system-oriented tasks in organizations, coping with the lack of adaptability and inter-operability in workflow systems is a manual and resource intensive effort. Recent advances in (i) standardized, modular, and distributed software delivery frameworks, such as web services; (ii) standard semantic markup languages for developing domain ontologies; and (iii) intelligent process coordination frameworks promise the development of flexible, responsive, integrated, and organic process management architectures. In this paper, we present an intelligent mediator-based architecture for enabling EAI. Intraorganisational information sources and services are made available via a web-services,framework. An Integrated Service Planning and Execution (lSP&E) framework interleaves service composition and execution at the mediator to fulfill service requests. Processes that interleave information gathering and transactional tasks are generated using domain-independent Hierarchical Task Network (HTN) AI planning and a domain-specific ontology and then executed in a scalable and reliable manner. Benefits and limitations of mediator-based frameworks for EAI and topics for further research are discussed based on a prototype development experience.