移动数据库视图更新算法研究

由于移动无线网络的不稳定特性,有必要在移动主机端缓存数据。在移动数据库应用中,数据缓存采取实体化视图形式,由位于固定网络中的视图管理器来维护。为了减少无线传输量,使用视图增量更新算法。但是当网络带宽不足时,只能选择一部分增量进行更新。于是有人提出了根据动态的数据新鲜度优先级进行排队的视图更新算法PIU算法,但是这种算法仅仅考虑了视图整体数据的动态新旧度,即平均时间因素,所以仍存在许多缺陷。为了弥补这些缺陷,在这基础上,引入用户的访问概率和数据的复杂度两个因素,提出改进的PIU算法,有效地改善移动数据库视图更新的面向用户性、实时性和一致性。

Research on Algorithm of View Update in Mobile Database

In mobile computing,instability such as disconnection and weak connectivity of network,source scarcity of mobile host and asymmetry of network communication etc.makes caching data on mobile host become necessary.In the context of mobile database applications,data cached can take the form of a materialized view maintained by a version mechanism called view holder within the fixed network.In order to reduce the quantity of transmission on wireless network,we may use an incremental update algorithm.But when there is lack of bandwidth,only a part of the increment should be chosen to update.So a view update algorithm called PIU algorithm which grounds on the dynamic data freshness priority is presented.But this algorithm only takes dynamic new-old extent of whole view data into account,thus there still exists a lot of limitations.Base on this algorithm,an improved PIU algorithm which considers two other factors of user's access probability and data complexity is presented to fetch up these shortages.And it effectively improves the user-oriented,real-time and consistency problem in view update of mobile database.