We present a randomized benchmark generator for attesting the correctness and performance of online first-order monitors. The benchmark generator consists of three components: a stream generator, a stream replayer, and a monitoring oracle. The stream generator produces random event streams that conform to user-defined characteristics such as event frequencies and distributions of the events’ parameters. The stream replayer reproduces event streams in real time at a user-defined velocity. By varying the stream characteristics and velocity, one can analyze their impact on the monitor’s performance. The monitoring oracle provides the expected result of monitoring the generated streams against metric first-order regular specifications. The specification languages supported by most existing monitors are either a subset of or share a large common fragment with the oracle’s language. Thus, we envision that our benchmark generator will be used as a standard correctness and performance testing tool for online monitors.
CITATION STYLE
Krstić, S., & Schneider, J. (2020). A Benchmark Generator for Online First-Order Monitoring. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 12399 LNCS, pp. 482–494). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-60508-7_27
Mendeley helps you to discover research relevant for your work.