摘要
把分支机构的数据库作为出版者和分发者,把上级中心的数据库作为订阅者,首先在分发者数据库中对表建立快照代理、日志阅读代理、分发代理,和存储过程的复制事物,并在数据库中记录同步状态信息,然后分发者连接出版者,阅读出版物的事务日志,搜索出带有复制标志的INSERT、UPDATE、DELETE语句和其他更新事务提交给订阅者,当分支机构数据发生更新,日志读取代理即时将更新信息推到上级单位中心的数据库中,实现分发者和订阅者数据库数据两者数据的一致性,从而实现分布式数据库数据的同步更新。
The branch database is regarded as Publisher or Distribute, and the database of superviser center as Subscription. First in Publisher database,we establish snapshot agency, journal-read agency,distribute agency and storage of process, record synchronous-state information in database, then we connect Publisher to Subscription, read the journal of transaction of Subscription, search for Copy-Brand with the INSERT , UPDATE and DELETE sentence and other transactions, submit them all to Subscription. When the data of branch database occur updating, push the updated-information into the database of Subscription immediately, and the consistency of Publisher and Subscription database's data gets realized.
出处
《佛山科学技术学院学报(自然科学版)》
CAS
2005年第2期48-51,共4页
Journal of Foshan University(Natural Science Edition)