数据流滑动窗口聚集查询降载策略研究

Load shedding for sliding window aggregation queries over data streams

滑动窗口聚集查询在数据流管理系统中应用广泛,数据流到达高峰期,必须考虑滑动窗口聚集查询中出现的降载问题。分析了子集模型的特点和已有降载策略的不足,给出了数据流滑动窗口聚集查询降载问题的约束条件,提出了能保证子集结果产生的基于丢弃窗口更新策略的降载算法。理论分析和实验结果表明,该算法对数据流滑动窗口聚集查询降载问题的处理具有较高的有效性和实用性。

Aggregation queries with sliding window are widely used in data stream management system. Load shedding must be taken into account as data stream burst into the aggregation queries. This paper analyzed characteristics of subset model and deficiencies of current load shedding methods. Gave restrictions of the load shedding problem, and a load shedding algorithm based on the strategy of drop window update. It could guarantee the produce of subset result. The theoretical analysis and experiments show that the algorithm is effective and efficient for the load shedding of aggregation queries over data streams.