多数据库中间件的模型研究

Research on multi-database middleware model

提出一种新的多数据库中间件架构,使得异源独立分布数据的“上推/受访/集成”模型化,从而使异源独立分布数据的存取/复用/集成,能像单一关系数据库那样走向模型化,建立“DBMS的DBMS”;讨论了该模型中的几个主要关键技术,如数据的一致性、数据的复制策略、局部数据源的即插即用研究、缓冲区系统中的数据名空间。

A new multi-database middleware model was proposed, which supported the dataupload/integration/access modeling and these data were heterogeneous source and distributed independently. Thus the modeling of access/reuse/integration of data were made, which were heterogeneous source and distributed independently, just like single relational database, and then the DBMS over DBMS was built. Severalkey techniques in the model were discussed, such as data consistency, data copy stragety, plug and play of local data source, and data namespace problem in DHDB system.