Monolithic to Microservices + Docker = SDLC on Steroids!
2 likes364 views
The document discusses the transition from monolithic architecture to microservices, highlighting the adoption of Docker and a polyglot tech stack at SS&C Eze to enhance the software development lifecycle. It emphasizes improvements in developer experience, quick onboarding, and the effectiveness of testing and deployment strategies, such as branch by abstraction and feature toggles. Key takeaways include the necessity for strict engineering discipline and the advantages of testing services in isolation for more efficient delivery.
Monolithic to Microservices + Docker = SDLC on Steroids!
- 1. Ashish Sharma Monolithic to Microservices + Docker = SDLC on Steroids Director SWE, SS&C Eze. @asharma_im
- 2. About SS&C Eze 1,900Asset managers served worldwide 25years delivering investment management solutions $40m+Annual R&D Investment 1050+ Employees 44% Service 36% R&D
- 3. Eze Eclipse Data Integration Prime Brokers Administrators Trade Execution High-Touch Desk Algorithms Outsourced Trading Desks
- 5. Tech Stack
- 6. Idea! Innovators Let's build a cloud native Investment Suite!
- 7. Tech Stack 1.0 ● Front End SPA in Angular ● ~12 Backend services in C# ● RMQ & Nginx based messaging ● MSSql Server for storage
- 10. Eclipse Infrastructure Uniform Packaging Identical multistage environments TEST PERF UAT PROD Dynamic scaling Simplified CI/CD tool chain
- 12. Standardization Language/Framework agnostic standards and patterns Bootstrapping ● Secrets management ● Configuration Resource Access ● DB/Cache Access ● Authentication ● Authorization ● Logging and Tracing APIs Boundary ● API Standards ● Swagger ● Contract driven development
- 14. Come to office Download Code Build Restore DB/Register Services/Configuration Work on Story Monolithic Developer Experience: Less than Ideal
- 15. Come to office Download Code docker-compose up Work on Story Microservices Developer Experience
- 17. Self Contained Code ● Mono repo to micro repos ● Explicit dependencies, build tools, tests and packaging ● It works the same way in local, test, UAT and PROD
- 18. • Quick to onboard new developers • Massive increase in cross team pull requests! Onboarding
- 19. • Quick to onboard new developers • Massive increase in cross team pull requests! Onboarding What happened Team- A?
- 20. • Quick to onboard new developers • Massive increase in cross team pull requests! Onboarding Team-A contributing into Team-B’s repo
- 21. Testing
- 22. Service Tests ● Setup dependent services and data ● Environment reliability ● Slow ● Cleanup Resource-1…………………….Resource-n Service Tests Unit Tests Wouldn’t it be nice if service tests were as easy as unit tests
- 23. Test Data ● Web crawler inspired technique to download snapshots from immediate dependencies ● Snapshots are served as mock during test run ● Snapshots are checked into source control
- 24. Test Suite Mock DB Test Client Service
- 26. Easy to write Fast Consistent Coverage Report Cleanup Triggers on every check-in Unit Tests Service Tests Service Tests Now as Easy as Unit Tests
- 27. Chaos Tests MockTest Client Service X-Request-Id = “/price return 500” Propagate headers If X-Request-Id == “/price return 500” → set res.status = 500
- 28. Delivery
- 29. Delivery Weekend Deployments Extended Downtime Painful Rollbacks Multiple Deployments per day Zero Down Time Easy Rollbacks
- 30. Deployments
- 31. Patterns for Safe Delivery Branch by abstraction ● Dark launch ● Feature Toggles V1 V2 Clone Input Workflow Compare Results Performance stats V1 V2 Get Toggle Workflow ONOff
- 32. KEY TAKEAWAYS Microservices with polyglot stacks are magical but requires strict engineering discipline Invest into making developer experience quick and smooth Shift Left. Test services in Isolation Apply standard patterns for Safe Delivery
- 33. Thank you!