Tulsa: A Tool for Transforming UML to Layered Queueing Networks for Performance Analysis of Data Intensive Applications
Motivated by the problem of detecting software performance anti-patterns in data-intensive applications (DIAs), we present a tool, Tulsa, for transforming software architecture models specified through UML into Layered Queueing Networks (LQNs), which are analytical performance models used to capture contention across multiple software layers. In particular, we generalize an existing transformation based on the Epsilon framework to generate LQNs from UML models annotated with the DICE profile, which extends UML to modelling DIAs based on technologies such as Apache Storm.
- 1.Dubois, D.J., et al.: Model-driven application refactoring to minimize deployment costs in preemptible cloud resources. In: CLOUD 2016. IEEE Press, USA (2016)Google Scholar
- 2.Casale, G., Ardagna, D., Artac, M., et al.: DICE: quality-driven development of data-intensive cloud applications. In: MiSE 2015, pp. 78–83. IEEE Press (2015)Google Scholar
- 4.D2.1 Design and quality abstractions. http://www.dice-h2020.eu/deliverables/
- 5.UML Profile for MARTE: Modeling and Analysis of Real-Time Embedded Systems, Version 1.1, Object Management Group (2011)Google Scholar
- 6.Altamimi, T., Zargari, M.H., Petriu, D., Performance analysis roundtrip: automatic generation of performance models and results feedback using cross-model trace links, In: CASCON 2016, Toronto, Canada. ACM Press (2016)Google Scholar
- 7.D2.3 Deployment abstractions. http://www.dice-h2020.eu/deliverables/