In a Data Stream Management System (DSMS), continuous queries (CQs) registered by different applications inherently have different levels of importance (i.e., quality of service (QoS) and quality of data (QoD) requirements). Moreover, the shift to cloud services and the growing need for monitoring applications will inevitably lead to the establishment of data stream management cloud services. In such a multi-tenant environment, it is expected that the different applications/users will demand different QoS and QoD requirements for their CQs. In this work, we are proposing an approach that exploits the synergy between scheduling and load shedding to effectively handle different ranks of CQ classes, each associated with different QoS and QoD requirements. The proposed approach was implemented and evaluated within AQSIOS, our DSMS prototype.
展开▼
