Ce diaporama a bien été signalé.
Nous utilisons votre profil LinkedIn et vos données d’activité pour vous proposer des publicités personnalisées et pertinentes. Vous pouvez changer vos préférences de publicités à tout moment.

Paraiso cloud2016.key

Model-Driven Management of Docker Containers. Docker Containers Management with Model.

  • Identifiez-vous pour voir les commentaires

Paraiso cloud2016.key

  1. 1. University of Lille & Inria Lille - Nord Europe (France) Fawaz Paraïso | Stéphanie Challita | Yahya Al-Dhuraibi | Philippe Merle 9th International Conference on Cloud Computing(CLOUD 2016) Model-Driven Management of Docker Containers
  2. 2. Click to edit Master subtitle style Model-Driven Management of Docker Containers 1 | Context
  3. 3. Jun 27- July 2, 2016 . San Francisco Containers Technology 3 Containers add new efficiency to Cloud Computing Docker is the well know containers technology Docker Adoption Is Not Without Concern … Portability low resource consumption Lightweight
  4. 4. Click to edit Master subtitle style Model-Driven Management of Docker Containers 2 | Challenges for Managing Docker Containers
  5. 5. Jun 27- July 2, 2016 . San Francisco A distributed Event Processing Application 5 Word Word Count Count Count Count Rolling Rolling Ranking Rolling Ranking Rolling
  6. 6. Jun 27- July 2, 2016 . San Francisco Deployability of Application 6 Deployment Corrections This operation can be repeated severals time Resource consumed Time consumed Application Runtime Appli Currently, with Docker there is no way to guarantee the deployability of application
  7. 7. Jun 27- July 2, 2016 . San Francisco Challenges for Managing Docker Containers 1. Challenge 1: Lack of the verification 7
  8. 8. Jun 27- July 2, 2016 . San Francisco Resources Management at Runtime 8 Runtime App cpu memory …
  9. 9. Jun 27- July 2, 2016 . San Francisco Challenges for Managing Docker Containers 1. Lack of the verification 2. Challenge 2: Resource management at runtime 9
  10. 10. Jun 27- July 2, 2016 . San Francisco Synchronization between Design Time and Runtime 10 Runtime App Design Time
  11. 11. Jun 27- July 2, 2016 . San Francisco Challenges for Managing Docker Containers 1. Lack of the verification 2. Resources management at runtime 3. Challenge 3: Synchronization between design time and runtime 11
  12. 12. Jun 27- July 2, 2016 . San Francisco Inconsistency Use of Containers Across Organizations 12 The problem of: Maintainability remains Inconsistency Swarm Compose
  13. 13. Jun 27- July 2, 2016 . San Francisco Challenges for Managing Docker Containers 1. Lack of the verification 2. Resources management at runtime 3. Synchronization between design time and runtime 4. Challenge 4: Inconsistency use of containers across organizations 13
  14. 14. Jun 27- July 2, 2016 . San Francisco Outline 14 1.Context 2.Problem Statement 3.Our approach 4.Evaluation 5.Conclusion 6. Future Work
  15. 15. Click to edit Master subtitle style Model-Driven Management of Docker Containers 3 | Our Approach
  16. 16. Jun 27- July 2, 2016 . San Francisco Model-Driven Engineering (MDE) 16 MDE = Abstraction + Automation Model is an abstraction that enables to deal with complex reality in simplified way Automation is a model transformation that mutates on model to another Viewpoint V2 M2 Viewpoint V1 M1 Object Constraint Language (OCL)
  17. 17. Jun 27- July 2, 2016 . San Francisco Our Approach 17 Model-Driven Management of Docker Containers Our solution is four phase model-driven tool 1. Analysis of consistency 2. Resource management at runtime 2. Synchronization of designed and deployed Containers 3. Design Tool
  18. 18. Jun 27- July 2, 2016 . San Francisco Architecture Overview 18 Docker Model Runtime Legend: Docker artifact Model element Connector Generate artifacts Update model elements Connector
  19. 19. Jun 27- July 2, 2016 . San Francisco Docker Model (1/2) ‣ In MDE everything is a model ➡Container, Link, Host,Volume, etc… • All modeling artefacts in MDE are interrelated 19 Transformations DockerFile, compose, swarm Validation Design
  20. 20. Jun 27- July 2, 2016 . San Francisco Docker Model (2/2) 20 }[Merle et al. Cloud 2015] {
  21. 21. Jun 27- July 2, 2016 . San Francisco Docker Container Architectural Constraints 21 Container-a Container-b This architecture will fail during its deployment due to the cycle between containers context Container inv NoCycleBetweenContainerLinks: links->select(oclIsTypeOf(Link)).target->closure (links->select(oclIsTypeOf(Link)).target)->excludes(self) 
 OCL constraint
  22. 22. Jun 27- July 2, 2016 . San Francisco Synchronization 22 Docker Model Runtime Connector Two way synchronization
  23. 23. Jun 27- July 2, 2016 . San Francisco Connecting Docker Model Online 23 Docker Model Runtime Detect changes Connector DESIGN TIMERUNTIME Model Model update Change Detection Monitoring
  24. 24. Jun 27- July 2, 2016 . San Francisco Design Tool 24 Available here: http://tinyurl.com/occiware
  25. 25. Click to edit Master subtitle style Model-Driven Management of Docker Containers 4 | Evaluation
  26. 26. Jun 27- July 2, 2016 . San Francisco Use Case 26 Storm topology ranking bolt rolling count rolling count count count count word spout 1 word spout 2 20 VMs with 8 Docker containers deployed on a private Cloud
  27. 27. Jun 27- July 2, 2016 . San Francisco Overhead 27 Create action Avg. start time Docker Model Overhead Docker 168.509 sec - Docker with Model 170.382 sec 1,11 % Start action Avg. start time Docker Model Overhead Docker 5.033 sec - Docker with Model 5.04 sec 2,1 % Stop action Avg. start time Docker Model Overhead Docker 84.12 sec - Docker with Model 86.01 sec 2,25 % Container creation time Container starting time Container stopping time The experiment shows that the overhead introduced by Docker Model is negligible regarding all advantages provided by our approach 8 containers were used to perform this evaluation
  28. 28. Jun 27- July 2, 2016 . San Francisco Scalability of Docker Model at Runtime 28 Total of 50,000 containers where generated. For each generation, the model performs creation and update actions. The average time taken to generate 50 containers is 14.30 seconds
  29. 29. Click to edit Master subtitle style Model-Driven Management of Docker Containers 5 | Conclusion
  30. 30. Jun 27- July 2, 2016 . San Francisco Conclusion Presented an ongoing Model-Driven Management of Docker Containers MDE tooling enables to design, reason, and deploy containers Our approach provides to the users the possibility to verify the deployability of Docker containers at design time Tested our approach on a representative application scenario and deployed it to cloud 30
  31. 31. Click to edit Master subtitle style Model-Driven Management of Docker Containers 6 | Future Work
  32. 32. Jun 27- July 2, 2016 . San Francisco Future Work Investigate the adaptation of managed containers as a set of atomic changes Extended our approach to other container solutions 32
  33. 33. Questions & comments University of Lille & Inria Lille - Nord Europe (France) Thank you! http://occiware.org Fawaz Paraïso : fawaz.paraiso@inria.fr

×