| The emerging application-to-application paradigm allows an organization's data to be shared across applications in both intra-enterprise and inter-enterprise contexts. In order to seamlessly integrate information from disparate, heterogeneous, and multi-organizational applications and quickly adapt to new application contexts, rigid information modeling and proprietary data representation technology have proven inadequate. Conventional data integration approaches are based on database systems that represent data semantics using static structured data schemas. They cannot provide adequate facilities to satisfy the increasing need for effective and efficient application data integration from heterogeneous data sources. Applications need to be built on an information model that is standards-based, flexible, and adaptable. This research proposes a framework that adopts a semi-structured data integration approach using a data view mechanism and XML and XPath standards. A prototype of the framework is designed and implemented using a loosely-coupled mediator-agent-adapter architecture based on distributed component technology. An XML query language provides the basis for data view definition as well as query specification query language are defined here to provide a formal basis for implementation of query processing and logical query optimization. The mapping between the to facilitate implementation of query processing. The XMark benchmark for XML databases is used to assess the functionality and performance of the prototyped implementation of the query language and system. The benchmarking result and analysis establish a base line query performance measure for the proposed data integration framework and identify the characteristics of different system implementations. |
