【作者】 洪筱菡；

【导师】 朱跃龙；

【作者基本信息】 河海大学 ， 计算机应用技术， 2003， 硕士

【摘要】 数据集成是企业之间或企业内各部门协同合作的需要。它的目标是实现各个异构数据源之间的数据共享，从而有效地利用资源，提高整个应用系统的性能。 目前随着计算机技术特别是计算机网络技术的发展，传统的数据集成技术(如多库方法)已无法满足人们获取更多更新数据的需要。人们要求数据集成系统不仅能集成数据库系统中的数据，而且能集成非数据库系统中的数据；不仅能集成传统的结构化数据，而且能集成半结构化甚至非结构化数据，这是传统的数据集成技术难以实现的。 本文针对上述背景，首先讨论了异构数据集成的相关理论，并分析了其中存在的问题，然后分别从中间件方面以及XML在数据模型角度和数据交换规范角度阐述了基于XML构建数据集成中间件的可行性。最后在此基础上给出了一个基于XML的数据集成中间件模型，并且着重阐述了在此中间件模型中如何基于XML实现异构数据集成的视图定义和查询处理。 XML本身具有的与平台无关、易于扩展、交互性好、语义性强等特性，使得基于XML的数据集成中间件模型能够较容易地实现对各数据源的描述以及数据源之间的数据转换，从而解决目前异构数据集成中存在的联邦查询优化欠缺、以及对基于Web的异质数据源集成的扩展性较差等问题。更多还原

【Abstract】 Data integration is the demand of coordination between not only different enterprises but different departments inside them, and the goal is to realize data sharing between heterogeneous data sources, so as to utilize resources effectively and attain high performance of the whole application system.Today with the development of the technology for the computer especially computer network, traditional technology of data integration, such as multidatabase technology, can not meet the demand of applications for mass datas and new kinds of datas. New technology is needed in today’s data integration system, which can integrate datas not only from database systems but also from other systems, and can integrate not only the traditional structure datas but also semi-structure even un-structure datas.In this paper, against above backgrounds, we first introduce the state of research for data integration and problems, and then analyze the feasibility of heterogeneous data integration based on XML from the view of middleware and XML technology in both data model and data exchange specification. Finally, based on these analyses, we put forward a model of data integration middleware and stress the problem of how to realize the view definition for heterogeneous database systems based on XML and the problem of query processing.The character for XML showed in this model, which includes platform independence, easy for expansion, better interaction, abundance semantic, make the middleware model easy for the realization of description and transform for heterogeneous data sources in the field of data integration, and solve the problems of scarcity for union query optimization and poor ability for integration expansion with heterogeneous data sources based on Web perfectly.更多还原