一种数据流上截止期敏感的滑动窗口处理策略

在数据流上定义了截止期作为连续查询的实时约束，并建立了基于截止期的滑动窗口处理模型；提出了一种基于预测的截止期敏感的滑动窗口处理策略；在多滑动窗口查询处理环境中，提出了一种（k,m ）截止期约束机制，在提高查询实时性的前提下，尽可能的满足不同查询对结果丢失率的不同约束。实验结果表明该处理策略能够有效提高数据流上滑动窗口查询的截止期满足率。

A Deadline-Sensitive Sliding Window Processing Strategy over Data Streams

Deadline is defined as a real-time constraint of continuous query over data stream, and a deadline-based sliding window processing model is built. Also, a deadline-sensitive processing strategy based on prediction is proposed. Under the circumstance of multi-query, a (k,m deadline constraint is dedicated to meet different result-loss-ratio constraints on different queries with the premise of improvement of real-time performance. Experimental results show that this strategy effectively improves the deadline satisfaction ratio of sliding window processing over data streams.