【作者】 郑娅峰；

【导师】 鱼滨；

【作者基本信息】 西北大学 ， 计算机应用技术， 2005， 硕士

【摘要】 异构数据集成的任务就是通过给用户提供一个统一的应用平台,从而屏蔽底层数据源的不同,使得用户可以无缝而且灵活的访问这些异构的数据源。 本文研究了异构数据集成的相关理论和技术,对现有数据集成方法进行分析,采用Mediator/Wrapper方式并结合本体技术,提出了一种数据集成中间件体系结构——HDSQ(Heterogeneous Data Source Query)。该结构采用成熟的XML技术,并利用本体描述精确语义的优势,使用本体描述语言OWL描述全局数据模式,解决数据源的语义异构问题。同时采用Web Service技术,实现对各类数据源的封装和远程调用,使集成结构更适用于分布式环境。 基于上述理论,设计实现了一个数据集成检索系统,该系统提供数据源配置,数据源包装和查询处理等多项功能,通过统一界面提供对各种关系数据库和XML文档的透明访问。 本文的主要研究工作如下: (1) 使用本体(ontology)描述领域概念,构建了全局本体和局部本体及局部到全局的映射关系,解决了异构数据集成中语义冲突的问题。 (2) 设计了异构数据集成的体系结构,给出了各模块的具体功能,并实现了一个原型系统。 (3) 将各数据源包装器封装为Web Service,简化了调用机制,增加了系统的灵活性和可扩展性。更多还原

【Abstract】 The purpose of data integration system is to provide users with seamless and flexible access to heterogeneous data source through a unified query interface which can hide the difference of underlying data source.The paper focuses on correlative theories and technologies of heterogeneous data integration and presents HDSQ (Heterogeneous Data Source Query) architecture of data integration which is based on analysis and comparing the existing data integration approach and the character of architecture. HDSQ adopts Mediator/Wrapper method and combines with ontology technology. This architecture uses OWL (Ontology Web Language) to represent global schema.lt provides an effective approach to solve the problems of semantic heterogeneity on data integration. The HDSQ that implements the encapsulation and invoking of the data source based on Web Service technology is fit for the distributed system environment.Finally, a system prototype which is based on the methods and technologies for data integration and query is introduced. The prototype can completely perform integration of schema, query processing and source configure. It provides transparent access to existing database management systems and XML data through a unified query interface.The main works of this paper are as follows:(1) The concept of field is described by ontology. The global ontology and local ontology are constructed. The relation between them is presented.Themethod solves the problems of semantic heterogeneity on data integration.(2) The heterogeneous data integration architecture is designed The functions of components are described and the prototype system is realized.(3) The method that data source wrapper is encapsulated as Web Service makes invocation mechanism simple and improves the flexibility and the extensibility.更多还原