Multi-Query Optimization of Incrementally Evaluated Sliding-Window Aggregations

Online analytics, in most advanced scientific, business, and social media applications, rely heavily on the efficient execution of large numbers of Aggregate Continuous Queries (ACQs). ACQs continuously aggregate streaming data and periodically produce results such as max or average over a given window of the latest data. It has been shown that it is beneficial to use Incremental Evaluation (IE) for re-using calculations performed over parts of the ACQ window, and to share them in multi-query (MQ) environments among certain sets of ACQs. In this work, we re-examine how the principle of sharing is applied in IE techniques as well as in MQ optimizers. We provide an extensive taxonomy of IE techniques and a new approach of using the state-of-the-art IE techniques as part of MQ optimizers in a way that reduces the execution plan costs by up to 270,000x. We evaluate all of our solutions both theoretically and experimentally using both real and synthetic datasets.