Abstract: A preemptive rate-based batching approach is proposed based on the characters of the data stream. A query plan is a sequence of operators. Operators in a query plan are divided into different operator units, which are given different priorities in the scheduling strategy. To improve the efficiency of the continuous query, the approach is preemptive and operator units’ priorities are changed while the system factors alter. Some experiments prove that the approach can not only improve the system’s performance, but also decrease the average latency time of tuples and improve the output rate of tuples greatly.

Key words: Rate-based, Batching, Preemptive

摘要： 针对数据流的特征，提出了一种基于速率的抢占式批处理方法。一个查询计划是一个操作符序列。文章将一个查询计划划分为不同的操作单元，并为不同的操作单元分配不同的优先级，而且这个优先级随系统因素的变化而动态改变，根据变化的优先级来动态调度操作单元，采取抢占式调度，从而提高连续查询的查询效率。实验表明该方法不但能提高系统的总体性能，而且可以减少元组的平均等待时间，大大提高了元组的输出速率。

关键词: 基于速率, 批处理, 抢占