数据流上加权共享滑动窗口的连接查询处理算法

Processing Shared Weight Sliding Window Join on Data Streams

在数据流应用中,系统经常需要处理大量的滑动窗口连续查询,采用共享滑动窗口技术可以有效节省存储空间,提高系统整体的查询处理能力。但是共享滑动窗口技术会增大单个查询的响应延迟,降低单个查询的服务质量。针对这个问题,论文提出了加权共享滑动窗口的概念,并提出了三种优化的连接执行算法,优先响应重要的滑动窗口查询,从而提高了系统整体的服务质量。理论分析和实验结果表明论文提出的方法是行之有效的。

In many data stream applications,there is a large amount of sliding window continuous queries need to be processed synchronously.The processing technique based on shared sliding window can save the memory efficiently, thereby improve the whole system processing performance.However,one disadvantage of this technique is that it increases the response time for each sliding window query.To address this problem,this paper proposes the concept of shared weight sliding window,and three join algorithms on shared weight sliding window,which can response the important sliding window query firstly,and consequently improve the QoS of the processing system.Both the theoretical analysis and experimental results show that the methods axe effective.