Successfully reported this slideshow.
Your SlideShare is downloading. ×

Rapid Performance Modeling by transforming Use Case Maps to Palladio Component Models

Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad

Check these out next

1 of 19 Ad

More Related Content

Slideshows for you (20)

Advertisement

Similar to Rapid Performance Modeling by transforming Use Case Maps to Palladio Component Models (20)

More from Heiko Koziolek (12)

Advertisement

Recently uploaded (20)

Rapid Performance Modeling by transforming Use Case Maps to Palladio Component Models

  1. 1. Rapid Performance Modeling by transforming Use Case Maps to Palladio Component Models Christian Vogel (KIT), Heiko Koziolek (ABB), Thomas Goldschmidt (ABB), Erik Burger (KIT) © ABB Group Slide 1
  2. 2. Motivation Performance Modeling at ABB  Industrial Automation Domain  Efficient control/data flow crucial for real-time constraints  Design sketches on whiteboards  End-to-end latency as a requirement  Challenges  How to reason on complex control/data flows?  How to break down global response time requirements to performance budgets for software components? © ABB Group Slide 2 Sketch of a simple webserver
  3. 3. Solution Approach UCM2PCM  Palladio Component Model (PCM) allows consideration and prediction of performance from the beginning  But  Control flow and structure in different views  PCM models are difficult to create for persons from industrial automation  Thus  Use Case Maps (UCM) as new frontend for PCM  Transformation from UCM to PCM © ABB Group Slide 3
  4. 4. Related Work Model transformations in SW Performance Engineering  Performance engineering  C. Smith and L.Williams, Addison-Wesley (2002) Performance Solutions: a practical guide to creating responsive, scalable software  D. Menasce, L.W. Dowdy, V. Almeida, Prentice Hall (2004) Performance by design: computer capacity planning by example  Survey of approaches for performance predictions  S. Balsamo, A. Di Marco, P. Inverardi, and M. Simeoni, IEEE TSE (2004) Model-based performance prediction in software development: A survey  H. Koziolek, Elsevier Performance Evaluation (2010) Performance evaluation of component-based software systems: A survey  Transformation of UCM into Layered Queuing Networks (LQN)  D. B. Petriu (2001) Layered software performance models constructed from use case map specifications  Use of UML Marte Profile for creating performance models  M. Woodside (2007) From annotated software designs (UML SPT/Marte) to model formalisms © ABB Group Slide 4
  5. 5. Foundations Use Case Maps  Scenario-based modeling language  Part of the User Requirements Notations (URN)  Intuitive and easy to learn  Graphical editor exists*  Allows specification of performance annotations © ABB Group Slide 5 * http://lotos.site.uottawa.ca/ucm/bin/view/ProjetSEG/WebHome
  6. 6. UCM2PCM Architecture © ABB Group Slide 6
  7. 7. UCM2PCM Performance Annotations added to UCMs © ABB Group Slide 7 Loop IterationCount Resource Demands Component AllocationBranch Probabilities Scenario Workload
  8. 8. UCM2PCM Mapping Approach © ABB Group Slide 8 UCM Path Component Responsibility
  9. 9. UCM2PCM Mapping Approach © ABB Group Slide 9 Return Call Component Call All calls
  10. 10. UCM2PCM Mapping Approach © ABB Group Slide 10 SystemRequiredCall SystemProvidedCall ParentCall
  11. 11. UCM2PCM Mapping Approach © ABB Group Slide 11 Usage Model
  12. 12. UCM2PCM Limitations  Limited Input Assistance  Performance Annotations need to be added as key-value pairs  Limited PCM feature support  Variables & Guarded Branches are not supported  No Customized ResourceEnvironment supported  Duplicated SEFFs  For every call of a function a new SEFF is created © ABB Group Slide 12
  13. 13. Evaluation Goals  Evaluation of Accuracy  Apply UCM2PCM to three test systems  Validate the correctness of the PCM models resulting from the UCM2PCM transformation  Measure and check the accuracy of the produced SimuCom performance results  Evaluation of Usability  Tutorial & Assignment for UCM2PCM  Survey among developers for getting feedback about the UCM2PCM tool © ABB Group Slide 13
  14. 14. Evaluation Tested Models  Media Store  Plain Java web application for storing and retrieving media files.  SPECjAppServer  Industry-standard benchmark, designed to measure the performance of application servers conforming to the Java EE 5.0 or later specifications.  Business Reporting System  Loosely modeled after a management information system, formerly analyzed at Carlton University. Users can retrieve live business data from the system and run statistical analyses. © ABB Group Slide 14
  15. 15. Evaluation Business Reporting System UCM © ABB Group Slide 15
  16. 16. Evaluation Accuracy: Original vs. transformed UCM model  Transformation successfully bridged gap between UCM and PCM  Simulation result difference to reference model below 15% and below 5% in most cases  Accuracy is seen sufficiently for early reasoning of different design alternatives © ABB Group Slide 16
  17. 17. Evaluation Usability: qualitative user survey  Main Advantages  Comprehensibility of model is high, also for non-experts  Fast(er) modeling  Modeling complex systems is seen critical with UCM2PCM  Results are only hints, without statistical relevance © ABB Group Slide 17
  18. 18. UCM2PCM Conclusions and Future Work  Transformation successfully bridges semantic gap between UCM & PCM  High result accuracy  Mainly positive feedback by survey  Future Work  Adding Input Assistance to UCM editor  Reverse transformation from PCM to UCM  Conducting further case studies & experiments © ABB Group Slide 18
  19. 19. © ABB Group April 25, 2013 | Slide 19

×