HoliAdapt——数据流中一种适应性查询处理策略  

HoliAdapt—An adaptive query processing strategy over data stream

作　　者：张立杰[1] 李志强[1] 宋宝燕[1] 

机构地区：[1]辽宁大学信息科学与技术学院

出　　处：《计算机应用》2006年第9期2028-2030,共3页journal of Computer Applications

基　　金：辽宁省2005年博士启动基金资助项目(20041029)

摘　　要：针对数据流上连续查询处理的特征,为适应流数据特征以及执行环境的易变因素,该文从选择率和执行时间的角度出发,考虑内存使用量和输出延迟适应性因素,提出一种适应性的查询处理策略—HoliAdapt。该策略基于查询窗口动态地收集统计信息,利用数学方法不断地优化查询计划,通过核心调度方法,对操作符进行适应性的调度。实验证明该策略有效地减少时间延迟和内存使用量,最终提高系统查询的效率。With regard to the properties of queries over data stream, we thought carefully about the adaptive factors： run-time memory usage and time latency based on selectivity and processing time in order to adapt changed factor of streaming data and execute environments. Then we proposed an adaptive query processing strategy - HoliAdapt. First, we dynamically collected statistical information over query windows, and then we continuously optimized query planning by mathematical means. Thereafter, we adopted a scheduling strategy to schedule operators in terms of adaptivity. Experiments have proved that this query processing strategy can effectively minimize run-time memory usage and time latency, and ultimately improve performance of query.

关 键 词：数据流 适应性 查询计划 调度 

分 类 号：TP311.131[自动化与计算机技术—计算机软件与理论]