CORE

Complex Event Recognition (CER) systems are a prominent technology for finding user-defined query patterns over large data streams in real time. CER evaluation is known to be computationally challenging, since it requires maintaining set of partial matches, and this quickly grows super-linearly the number processed events. We present CORE, novel COmplex event Engine that focuses on efficient class complex queries, including time windows as well partition-by correlation operator. This engine uses automaton-based algorithm circumvents super-linear match problem: under complexity, takes constant per input maintain structure compactly represents matches and, once found, results may enumerated from with output-linear delay. experimentally compare CORE against state-of-the-art real-world data. show (1) CORE's performance stable respect both window size, (2) outperforms other by up five orders magnitude different workloads.