【作者】 贾晖；

【导师】 吴江；

【作者基本信息】 西北大学 ， 计算机应用， 2005， 硕士

【摘要】 数据集成是为了消除企业内部信息孤岛以及实现企业之间信息共享,实现异构数据透明访问,是目前研究热点之一。本文提出并研究了异构数据集成中间件——HDIM,其作用是提供一个访问异构数据源的中间平台,使用户不必考虑数据模型的异构性、分布性,透明的访问异构数据源。 本文作为2004年度科技型中小企业技术创新基金《基于异构系统的电信级客户关系管理系统》核心技术预演。主要目的是解决数据源的系统异构,模式异构,位置异构,以及部分语义异构问题。 本文的主要工作如下: 1.设计了异构数据集成中间件体系结构:以典型的Mediator-wrapper体系结构与基于本体的异构数据集成体系结构相结合作为本文研究的数据集成中间件——HDIM的体系结构。 2.采用两种本体描述,两级本体映射的GCA方式实现全局本体到数据源的映射。两种本体是全局本体和局部本体。全局本体作为局部本体的统一视图,描述中间件在特定领域范围内可集成的所有数据。局部本体是对特定数据源的本体描述,是进行本体映射的基础。两级映射分别是全局本体到局部本体的映射,局部本体到数据源的映射。通过两种本体,两级映射实现了以全局本体为核心的(GCA)查询模式。 3.采用Web服务解决中间件与数据源的通讯问题,屏蔽了数据源的位置异构,实现数据源的增、删、改与中间件分离。 本文工作的意义在于采用新技术为解决异构数据集成问题提出新的解决思路,使用户及其它应用程序可以只通过访问全局本体就能得到领域内所有的相关数据,为数据共享,数据分析,决策支持等应用提供支持。更多还原

【Abstract】 In order to eliminate informational isolated island in enterprise and solve information share between enterprises, data integration become a hot point in research area. This paper research and develop on the heterogeneous data integration middleware. The role of the middleware is as a middle platform for information access. Users of the platform do not need consider the heterogeneity and distribution of data source.The research of this paper is as kernel technique of a project: Client Relationship Management in telecom enterprise based on heterogeneous system, which is supported by middle or small science and technology enterprise innovation fund. This paper try to solve system heterogeneous, model heterogeneous, position heterogeneous and part of semantic heterogeneous.The main contributions of this paper are as follows:1. Present heterogeneous data integration middleware system architecture: combine classical Mediator-wrapper system architecture and ontology-based heterogeneous data integration architecture.2. Use two kinds of ontology, two levels of ontology mapping. Two kinds of ontology are global ontology and local ontology. Global ontology is as a unified view of local ontology, describing all datum in specific domain that can be integrated in middleware. Local ontology is the ontology description of specific data source, is the basis of ontology mapping. Two levels of mapping is the map between global ontology and local ontology and the map between local ontology and data source. Through two kinds of ontology and two maps, middleware can form global ontology centric approach in data query.3. Use Web Services figure out communication between middleware and datasource. Make data source independent with middleware.The meaning of this paper’ s research is using new technology proposing new method to solve heterogeneous data integration. User or other application can access global ontology, through ontology mapping, and get all data in specific domain.更多还原