The project proposes to develop a meta-object model and integrate it into the design of workflow systems. The meta-object model combines the affinity (desire-to-know) and privilege/obligation (need-to-know) information into a unified information model which facilitates efficient information access and secure access control. The Adlets and Malleable objects in this meta model are active objects that capture inter-dependency among object accesses. This information model is to be used to develop a new methodology for flexible and dynamic workflow system design. Traditional workflow systems are either process- or product-based. Different from these existing systems, workflows in the proposed information-based system models all workflows as activities of information accesses. As a result, it allows for incomplete specification of the initial workflow and dynamic construction of the workflow, i.e., exception and special sub-workflows are generated as needed when information accesses are invoked. There are two significant technical impacts in the research: the active meta-object model supports both performance and security for information access in a single uniform model, and the incorporation of it presents a new approach for workflow design which is more flexible and dynamic.
The research project is based on the premise that the design of a workflow system can be made more flexible, dynamic, and secure through the integration of an information access model in workflow systems where additional state and semantic information of subjects and objects can be captured through access events. The collaborative research tasks consist of modeling of the information model (the University of Florida), its implementation support on overlay networks (the University of Pittsburgh), and a practical domain-specific application in building construction (School of Building Construction at the University of Florida). The research assumes a service-oriented computing paradigm, where workflows are considered as compositions of web services. The primary contribution of the research project is the development of an efficient service composition framework based on dynamic service substitution and adaptation during the design, binding, and execution phases of a workflow system. The architecture relies on a new concept called abstract web service, which represents a community of similar services. Service communities are clustering of logical services with various degree of "similarity" for the purpose of finding replacement for service substitution or adaptation. An automata-based similarity measurement methodology was developed as part of the service composition framework. Akin to the domain name service in internet, service composition in web requires a registry to facilitate the posting, searching and discovery of services. The design of the structure and management of service registries rely on efficient and precise similarity analysis. The logical design of the service composition framework and the physical support of the service discovery are two synergetic research components performed collaboratively between UF and UPitt. The significance and practicality of the research project is demonstrated in its application to building construction. This particular application domain exhibits large fragmentation in resources and information, and thus can benefit greatly from an information system that allows for effective integration of workflows. The research task leverages existing standards (e.g., Master format for building materials) or models (e.g., the BIM building information model) for extracting semantic information to enhance interoperability in building construction processes. The research accomplishments include the conceptualization of construction processes and construction objects, and a framework for sharing, exchanging, and integration of information involving the reconciliation of multiple data formats, structures, and schemas in construction workflows. All research components share a common ontology-based approach toward the overall goal of providing support for dynamic web service composition. The automation of a workflow or service composition involves sharing of documents and relevant resources between participants. Sharing and reuse of knowledge can be driven by perspective ontologies. In particular, with respect to enriching service discovery, ontology-based service specification facilitate semantic-aware service discovery. Also in the target building construction application, we rely on the development of ontologies for seamless integration of various phases of the building construction life cycle. More precisely, the research findings and accomplishments are listed as follow: Development of the integrated web service composition framework. Introduction of the concept of service community. Development of methodology for similarity measure based on automaton. Development of a structured service registry with service discovery algorithms. Implementation of network support for semantic-based web service discovery. In corporation of context-awareness into the composition framework and service registry. The educational training component of the research project includes the development of new graduate courses on service-oriented computing and context-aware computing, and the completion of six PhD dissertations (four at UF and two at UPitt).
