智能与分布计算实验室
  基于web服务的多数据库研究
姓名 刘钊
论文答辩日期 2004.05.09
论文提交日期 2004.05.10
论文级别 硕士
中文题名 基于web服务的多数据库研究
英文题名 Research of Multidatabase Based on Web Service
导师1 卢正鼎
导师2
中文关键词 分布式计算;多数据库系统;网络服务;简单对象访问协议
英文关键词 distributed computing;multidatabase system;Web Service;SOAP
中文文摘 随着数据库技术和网络技术的发展,人们对数据资源共享的要求越来越高。多数据库系统为有效地集成多个分布、异构和自治的数据库提供了很好的解决办法,从而满足人们对数据资源共享的要求。 传统的多数据库系统多采用基于通用对象请求代理框架的集成和通信方案,这样的方案在系统间通信时往往需要构建专门的通信组件来满足要求。在分析Web服务这种新型的分布式计算模型并将之与传统的分布式计算模型比较的基础上,对传统的多数据库系统结构进行了改进,给出基于Web服务的多数据库体系结构模型,并以此为基础探讨将多数据库的各项功能完备有效的封装成Web服务的具体方法。 在采用Web服务的多数据库框架的集成系统中,存在海量数据传输的效能瓶颈。基于此分析了Web服务底层通讯协议??简单对象访问协议。在大数据量传输的前提下,我们把基于简单对象访问协议的请求和回应分解为若干个阶段并对各个阶段进行行为分析,给出了“块编码”、“永久连接”和“基于完全哈希表匹配的消息接收”等措施来提高传输效率。同时也分析了传输控制协议协议承担的时耗并对此给出诸如“置客户端于主动关闭状态”和“设置传输控制协议的延迟确认为0”等措施,取得了不错的成效。 利用Web服务的思想和部分Web服务技术,将已存的多数据库系统Panorama的数据源从局域网络扩展至广域网络,使得Panorama的查询客户端不仅可以通过统一查询界面查询位于局域网络内部的数据源,还可以通过自行设计实现的Web服务访问位于局域网络以外的数据源。
英文文摘 With the rapid development of the network and database technology, the needs for data sharing is getting stronger than ever before. Multidatabase system has rendered a very good solution to the integration of several distributed, heterogeneous and autonomous databases.The multidatabase system meets the needs of data sharing by shilding the physical and logical differnecs between databases. Multidatabases ever existed almost take CORBA as the base of system and communication integration, which often meet with such obstacles as implementation of new components for communication between systems when integrated systems needs interoperation. Web Service is a service oriented distributed computing module. On the basis of analysis and comparing with different theories and methods in integrating heterogeneous database systems, a new distributed multidatabase archietecture is proposed, which has such strongpoints as loose-coupling, cross-platform, cross-programming language, can be extended to Internet easily. Some self-contained and effecitive methods to encapsulate sub-functions of multidatabase into web services are suggested. There exists bottle neck in the transmission of massive data in the integrated system based on Web Service. On the basis of analysis of SOAP sending message and receiving message processes, specific methods such as block-coded, persistant connection and message receiving based on perfect hash table are proposed and verified to be effecitive after analyzing the performance of SOAP which is the communication protocol of Web Service under massive data transmission. Analysis of TCP proves that TCP is some reason why the effectiveness of SOAP is low and some effective methods such as make active-closed client are proposed. An extended multidatabase prototype, named Panorama system, was designed and implemented partly with above methods. A detailed implementation of the local agent named WebLDA is introduced, which can extract data from databases from Internet, therefore provides transparent interoperability among web database sources and existing database management systems.