Reusable patterns for scalable APIs running on Docker

The shipping containers were introduced around 1830s and since then they play a crucial role in the modern society by providing efficient packaging, storage and transportation. Today we see the same revolution happening in the DevOps world with the adoption of container technologies like Docker. They allow us to package, distribute and run software in a scalable and efficient way. In this talk we’ll see how we can abstract the common problem we solve everyday when building scalable Java APIs with Docker into design patterns to create reusable solutions. We’ll explore the good practices of writing Java applications with Docker. Then we’ll see how each design pattern can be applied in real scenarios that address different concerns that a large system creates. We’ll see some real life implementations of those patterns and how they help us solve problems in scalable systems. By the end of the talk we’ll have a very powerful abstraction to tackle the everyday problems we face in building big and impactful systems.

Transcript

3. Containers
5. cgroups
6. namespaces
7. Copy-on-write
8. Stackable Image Layers
9. bootfs
10. bootfs Base image - (ubuntu, debian)
11. bootfs Base image - (ubuntu, debian) Add JRE
12. bootfs Base image - (ubuntu, debian) Add JRE Add git
14. https://store.docker.com/images/alpine
15. https://www.dockerbook.com/
16. Class Object Image Container
18. Map/Reduce
19. https://www.usenix.org/node/196347
20. Single-node, multi-container Application patterns
21. Sidecar Web server Log server File System
22. Sidecar Application Instance Translations Downloader File System
23. Resource Isolation Separate development Easy and flexible composition Failure containment
24. Ambassador Web Server Ambasador External App External App External App
25. Ambassador example Application Hyperbahn Instance Hyperbahn Instance Hyperbahn Instance Hyperbahn Instance https://github.com/uber-archive/hyperbahn
26. Easy to reason about Easy to test Reusable
27. Adapter Application Monitoring Adapter Centralized monitoring system
28. Multi-node application patterns
29. Leader election App Leader election container App Leader election container App Leader election container App Leader election container
30. Simplifies the problem Easy testing
31. Distributed Transactions App Transaction manager container App Transaction manager container App Transaction manager container App Transaction manager container
32. Work queue Worker Work Execution Worker Work Execution Work Coordinator
33. Simplification
34. Scatter/gather Merge Execution Work Execution Work Execution Root Work Execution
35. Sharded Service Shard Function Distributor Worker Shard 1 Worker Shard 2 Worker Shard 3
36. http://enterpriseintegrationpatterns.com
37. Messaging endpoints
38. Idempotent Receiver App Idempotent Receiver
39. Message Routing
40. Dynamic Router App Consumer App Consumer App Consumer Dynamic Router Output Channel Controll Channel
42. Thanks!