Abstract

The scheduling for hybrid query set composed of periodic, continuous and one-time queries with timing constraints is an important issue in data stream management systems. At first, a hybrid real-time query model including periodic and aperiodic queries is introduced. Moreover, a feedback control scheduling strategy, termed FC-TBS, is proposed to schedule these queries to guarantee the deadlines of periodic queries and minimize the number of deadline violations for aperiodic queries. At the same time, the FC-TBS tries to improve the overall query quality by adaptive adjustment on CPU utilization factor for aperiodic queries according to workload characteristics and application defined relationship between sample ratio and QoS. The experimental results show the adaptive scheduling strategy has a noticeable advantage in dealing with real-time queries on dynamic data streams.

Paper Details