数据流管理系统中Max、Min聚集算子的示例概要算法

Max和Min是数据流管理系统中重要聚集算子。应用基于滑动窗口下的示例概要法在实时数据流场景下计算Max和Min。在本方法中不需要保存所有落入滑动窗口中数据元组，这意味着可以极大地减小存储空间。由于存储元组的减少，系统的处理时间也显著地减少。实验结果表明基于滑动窗口的示例概要法显著降低了时间和空间的开销。

Calculating Max and Min with exemplary sketch algorithm in data stream management system

Max and Min are two important aggregates in the Data Stream Management System (DSMS). A novel mechanism called the exemplary sketch was proposed to evaluate Max and Min over sliding windows with various sizes in the real time data stream environment. With this method we no longer need to buffer all the tuples to fall into current sliding windows, which means we can reduce the space cost remarkably. Because of this, the processing time can also be reduced dramatically. Experimental results show that the sketch scheme yields very good performance on both space and time cost.