移动数据库中数据复制同步处理策略的研究

Research on synchronous processing strategy of data replication in mobile database

同步技术是保证复制移动数据库系统一致性的一项关键技术,鉴于目前移动复制同步技术存在通讯数据量大,存储空间消耗多,尤其是在网络带宽下降时,不能及时更新客户端的数据,导致移动事务执行失败等缺陷。通过UTLRSP(Union Transaction-Level Result-Set Propagation,关联事务结果集)复制同步模型结合数据广播技术,并利用基于优先级的增量更新算法实现客户端与中心数据库服务器的数据同步处理。实验结果表明,与两级复制机制相比,UTLRSP模型将事务作相关联处理,且只保存事务结果,有效降低了存贮空间的消耗,减小了同步过程中通讯数据量;基于优先级的增量更新算法根据数据新鲜度排列优先级,保证在无线网络带宽下降时新鲜度最高的数据先传输,提高了数据的传输效率、动态新鲜度以及客户端的可扩展性。

Synchronization is one of the key technologies in maintaining the consistency of replicated mobile database systems. The current mobile replication synchronization technology has many defects, such as the large amount of communication data and storage space consumption, especially, when the network bandwidth falls, it cannot update the client data timely, which lead to fail to perform mobile transactions. In order to solve the above shortcomings, the paper adopts UTLRSP（Union Transaction-Level Result-Set the Propagation）replication synchronization model combined with the data broadcasting technology, and by using the incremental updating algorithm based on priority to realize client and center database server data synchronization processing. Experimental results show that the model effectively reduces the consumption of client storage space and cuts down the communication costs compared with the two levels of replication mechanism.It associates the performing transaction with other transactions, and only saves transaction execution results. When the wireless network bandwidth falls, the algorithm prioritizes data according to the freshness. It ensures the highest freshness data transfer firstly. The algorithm improves the efficiency of data transmission, dynamic freshness and scalability for the client.