Research on Dynamic Integration of Heterogeneous Information Sources in Network Environments

近年来，随着计算机和网络技术的发展，异质信息源的集成已经成为一个非常重要的研究课题。在这项研究中，我们研究了分布式环境中信息源的动态集成的若干问题。第一个问题是用于数据集成的数据模型。第二个问题是基于中介器/包装器体系结构的信息集成系统中的数据处理。第三个问题是关于用户界面和基于代理的查询处理。主要研究成果总结如下：1.本文的主要研究成果。我们开发了一个名为WebNR/SD的数据模型。它侧重于关系数据库、XML文档库和WWW的动态集成。在WebNR/SD中，数据被表示为可能包含结构化文档值的嵌套关系。WebNR/SD代数允许在结构化文档和嵌套关系之间进行动态转换。设计了一个基于…的信息集成系统更多的中介器/包装器架构。我们实际上实现了一个名为InfoWeaver的原型系统。实现了Oracle、OpenText、XML WWW等页面的数据集成。在WebNR/SD中，用户可以查询驻留在远程WWW服务器上的WWW页面。提出了一种基于代理的WWW查询处理方案，并通过实验验证了该方案在降低页面传输成本方面的有效性。我们开发了一个可视化的用户界面，帮助初学者明确数据集成请求，并在InfoWever上进行了实现。我们开发了一个方案来实现视频数据库在我们的环境中的集成。使用SMIL来描述多媒体集成。我们开发了交互式WWW结构创作工具。支持将数据集成结果以多媒体WWW页面的形式发布。以传播为基础的信息传递服务已变得重要。我们将InfoWeaver扩展为具有基于传播的功能：基于传播的信息源的整合和基于传播的整合结果的交付。我们开发了一种可扩展的XML数据库查询语言。我们研究了与中介系统中查询以外的数据处理相关的问题。较少

Due to the recent advances of computer and network technology, integration of heterogeneous information sources has become a very important researchtopic. In this research, we have investigated a number of issues for dynamic integration of the information sources in distributed environments. The first issue is on a data model used for data integration. The second issue is on data processing in information integration systems based on the mediator/wrapper architecture. The third issue is on user interface and on agent-based query processing. The main research results are summarized as follows.1. We developed a data model named WebNR/SD. It focuses on dynamic integration of relational databases, XML document repositories, and WWW. In WebNR/SD, data is represented as nested relations possibly containing structured document values. The WebNR/SD algebra enables dynamic conversion between structured documents and nested relations.2. We designed an information integration system based on the … More mediator/wrapper architecture. We actually implemented a prototype system named InfoWeaver. It achieves integration of data in Oracle, OpenText, XML WWW pages, and so on.3. In WebNR/SD, users can query WWW pages residing on distant WWW servers. We developed an agent-based WWW query processing scheme, and verified its effectiveness on the reduction of page transfer cost through experiments.4. We developed a visual user interface which help novice users specify data integration requests, and implemented it on InfoWeaver.5. We developed a scheme to attain integration of video databases in our environment. SMIL is utilized to specify multimedia integration.6. We developed interactive WWW structure authoring facility. It supports publishing data integration results as multimedia WWW pages.7. Dissemination-based information delivery service has become important. We extended InfoWeaver to have dissemination-based features : integration of dissemination-based information sources and dissemination-based delivery of the integration results.8. We developed an extensible query language for XML databases.9. We studied issues related with data processing other than querying in mediation systems. Less