Service-oriented systems are designed for interconnectingwith other systems. The provided services face timing con-straints, the so-called latencies. We present a high-level per-formance evaluation technique that can be used by a systemdesigner to obtain distributions of these latencies. This tech-nique is capable of capturing nondeterministic, probabilisticand real-time aspects in one go. Under the hood, the tech-nique is equipped with two mechanisms: (i) selection of theright abstraction of the model (to prevent a state space ex-plosion) by evaluating the performance of executing modelsof different complexities; and (ii) an efficient algorithm inwhich basic estimates, simulation, and (probabilistic) modelchecking are combined. We illustrate our approach with ancase on image processing of interventional X-ray systems.
|Title of host publication||ModelEd, TestEd, TrustEd|
|Editors||Rom Langerak, Arend Rensink|
|Publication status||Published - 27 Sep 2017|
|Name||Lecture Notes in Computer Science|
van den Berg, F. G. B., Haverkort, B. R. H. M., & Hooman, J. (2017). iDSL: Automated Performance Evaluation of Service-Oriented Systems. In R. Langerak, & A. Rensink (Eds.), ModelEd, TestEd, TrustEd (Lecture Notes in Computer Science). Springer. https://doi.org/10.1007/978-3-319-68270-9_11