【摘要】 对分布式数据流进行查询,得到数值最大的K个对象(Top-K观测查询),最直接的解决方法是由中心结点处理分布式数据流,但这种方法导致中心结点和网络负载较大。提出一种基于动态修正值的查询算法,通过对观测数据进行计算得到修正值,并利用该修正值对不同结点处的对象数据进行操作,从而无需将结点数据流全部发送到中心结点就能完成Top-K观测查询。因而可以减少对网络带宽的要求和降低中心结点的负载,同时还能保持查询结果的完全准确。更多还原

【Abstract】 The K largest values can be obtained by querying the distributed data streams(Top-K monitoring queries).The straightforward method is to process the data streams on a coordinator.However,this leads to heavy loads on the coordinator and the network.A dynamic adjustment factor based algorithm for Top-K monitoring queries is proposed.The overall communication cost and the load of the coordinator can be reduced significantly,and at the same time the preciseness of the query results is fully guaranteed.更多还原