A Method for Aspect-Oriented Meta-Model Evolution. VAO 2014 York. Reiner Jung Robert Heinrich Eric Schmieders Misha Strittmatter Wilhelm Hasselbring

Transcription

1 for spect-oriented Meta-Model Evolution VO 2014 York Reiner Jung Robert Heinrich Eric Schmieders Misha Strittmatter Wilhelm Hasselbring 22 nd July 2014 iobserve This work was partially supported by the DFG (German Research Foundation) under the Priority Program SPP1593: Design For Future Managed Software Evolution (grant H 2038/4-1, RE 1674/6-1, PO 607/3-1) and the Helmholtz ssociation of German Research Centers

2 Motivation Introduction Driver of Meta-Model Evolution 1. Long-living software systems 2. Convergence of design-time and run-time models e.g. MPE scenarios [Ghezzi, 2011] spect-oriented Modeling (Concern-Driven Development) Separate concerns Reduce meta-model complexity Ease meta-model evolution and reuse Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

3 Example: Palladio Meta-Model [ecker, Koziolek, & Reussner, 2009] Introduction Palladio Domain Performance prediction Meta-Model 188 classes (PCM) Concerns s, deployment, workloads,... Scenario Reuse PCM as run-time meta-model for forecasting Extensions Cloud Monitoring Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

4 Related Meta-Modeling pproaches Related Work Meta-Model Extension by Subclassing [Steinberg et al., 2009] Decorator Pattern pproach EMF Profiles [Langer et al., 2012] Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

5 Meta-Model Extension by Subclassing [Steinberg et al., 2009] Related Work rchitecture Meta-Model Model Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

6 Meta-Model Extension by Subclassing [Steinberg et al., 2009] Related Work rchitecture Meta-Model Model Probed Monitoring MM Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

7 Meta-Model Extension by Subclassing [Steinberg et al., 2009] Related Work rchitecture Meta-Model Model Probed Monitoring MM Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

8 Meta-Model Extension by Subclassing [Steinberg et al., 2009] Related Work rchitecture Meta-Model Model Cloud Probed Cloud MM Monitoring MM Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

9 Meta-Model Extension by Subclassing [Steinberg et al., 2009] Related Work rchitecture Meta-Model Model Cloud Probed Cloud MM Monitoring MM Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

10 Solution Requirements cf. [Langer et al., 2012] 1. Non-invasive meta-model annotation mechanism 2. Separation of concerns from different modeling domains 3. Honor different types of relationships and roles of meta-models 4. Should support a formal specification for extensions/annotations 5. Must support existing technology to ease integration 6. Easy application and compatible with existing tooling Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

11 Overview Central ideas spect-oriented meta-models Semantic of references Contextual meta-model patterns Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

12 spect-oriented Modeling Roles rchitecture Meta-Model Model Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

13 spect-oriented Modeling Roles rchitecture Meta-Model Monitoring MM Model base aspect Probe Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

14 spect-oriented Modeling Roles rchitecture Meta-Model Monitoring MM Model Probe Deployment Meta-Model Model Computing Node Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

15 spect-oriented Modeling Roles rchitecture Meta-Model Monitoring MM Model aspect Probe Deployment Meta-Model Model base Computing Node Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

16 spect-oriented Modeling Roles rchitecture Meta-Model Monitoring MM Model Probe Deployment Meta-Model Cloud Meta-Model Model Computing Node base aspect Node Configuration Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

17 Semantic of References Syntactical View on References Semantical View on References Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

18 Semantic of References Syntactical View on References Semantical View on References Monitoring rchitecture Probe spect/ase Model Roles Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

19 Semantic of References Syntactical View on References Semantical View on References Monitoring rchitecture Probe spect/ase Model Roles Property ody Type Declaration Typing & Declarations Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

20 Semantic of References Syntactical View on References Semantical View on References Monitoring rchitecture Probe spect/ase Model Roles Property ody Type Declaration Typing & Declarations Queue Trace Derived Models Derived Node Original Node Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

21 Contextual Meta-Model Pattern Use Cases Generators Editors Simulation and Evaluation Pattern Tracability Navigation Queries Execution / ehavior Data State Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

22 Use Case: Run-Time Model Palladio Model <<component>> <<component>> Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

23 Use Case: Run-Time Model Observation Data Model Observation Update Transformation Palladio Model <<component>> <<component>> Monitoring Trace Model Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

24 Use Case: Run-Time Model Observation Data Model Observation Update Transformation Palladio Model <<component>> <<component>> Monitoring Trace Model PCM2LQN Transformation λ μ Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

25 Use Case: Run-Time Model Observation Data Model Observation Update Transformation Palladio Model <<component>> <<component>> Monitoring Trace Model PCM2LQN Transformation LQN2PCM Trace Model λ μ Evaluation Deployment daptation Result Model Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

26 Summary and Outlook Conclusions Summary Non-invasive annotation mechanism through OM Separate concerns represented by separate meta-models etter understanding of pattern and semantics in meta-models Use EMF/Ecore as formal notation for meta-models No new meta-meta-model required Compatible with existing frameworks Outlook Detailed definition of meta-model pattern Implications on tooling Evaluation of our method in a PCM modernization effort Solid core PCM Wide palette of extensions Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

27 ppendix Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

28 Decorator Pattern [Gamma, Helm, Johnson, & Vlissides, 1994] Conclusions ase Meta-Model Model bstract Decorator Cloud Reliable Cloud MM Reliability MM Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

29 EMF Profiles [Langer et al., 2012] Conclusions Features Provide a profile mechanism for EMF Multiple profiles can be annotated to one class Generic reusable profiles for reuse Comes with its own (meta-)meta-model Downside No tool-integration Xtext DSL framework [XText, 2011] KLighD diagram and auto-layout framework [Schneider, Spönemann, & von Hanxleden, 2013] Genmodel facilities Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

30 References I Conclusions G, I., & Foundation, E. [2011]. Xtext. ecker, S., Koziolek, H., & Reussner, R. [2009]. The Palladio component model for model-driven performance prediction. J. of Systems and Software, 82, Retrieved from doi: /j.jss Gamma, E., Helm, R., Johnson, R., & Vlissides, J. [1994]. Design patterns: elements of reusable object-oriented software. Pearson Education. Ghezzi, C. [2011, Sept]. The fading boundary between development time and run time. In Web services (ecows), th ieee european conf. on (p. 11). doi: /ECOWS Langer, P., Wieland, K., Wimmer, M., & Cabot, J. [2012]. EMF Profiles: lightweight extension approach for emf models. JOT, 11(1), Retrieved from Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12

31 References II Conclusions Schneider, C., Spönemann, M., & von Hanxleden, R. [2013, September]. Just model! Putting automatic synthesis of node-link-diagrams into practice. In Proceedings of vl/hcc. San Jose, C, US. Steinberg, D., udinsky, F., Paternostro, M., & Merks, E. [2009]. Emf: Eclipse modeling framework (2. ed.). oston, M: ddison-wesley. Retrieved from Jung, Heinrich, Schmieder, Strittmatter, Hasselbring Meta-Model Evolution / 12