Microservice Transformation of the Haufe Publishing System
The document discusses the Haufe Publishing System project and its approach using lean principles. Some key points: - The project aimed to modernize their platform by reducing dependencies, improving sharing, and allowing for faster change. - Lean principles like eliminating waste, amplifying learning, and deciding late were followed. Features were implemented incrementally and feedback was used for continuous improvement. - The architectural approach focused on business value, composability, shared services, evolutionary refinement, and data-driven processes. - The project used a pipeline approach with separate environments for development, integration, and production on Kubernetes clusters for each environment. Automated testing was done at each stage.
Microservice Transformation of the Haufe Publishing System
- 1. Haufe Publishing System Experiences from project HRS Universe 2020 Marco Seifried, Karsten Gaebert Workshop Barcelona May 10, 2017
- 3. Business Drivers • Running out of support • Speed of change • USP / Business Features • Reduce dependencies • (Tech) Sharing
- 5. Lean Principles ELIMINATE WASTE Speed! We only implement what‘s requested and release immediately. Get feedback. And start over. AMPLIFY LEARNING Learning is integral part of the project. DECIDE AS LATE AS POSSIBLE Make decisions based on facts Build in capacity for change DELIVER AS FAST AS POSSIBLE Design, implement, feedback, improve EMPOWER THE TEAM No fraction. Responsibility. Less control. BUILD INTEGRITY IN We deliver business value, in small doses that make sense SEE THE WHOLE Measure overall performance, not individual
- 6. BASELINE – ARCHITECTURAL PRINCIPLES Business value over technical strategy Strategic goals over project-specific benefits Composability over silos Shared services over specific-purpose implementations Evolutionary refinement over pursuit of initial perfection Design for obsoleteness over building for eternity Good enough over best of breed Declarative processes over implicit knowledge Data over opinion
- 7. Project Approach HRS 2020 EDB
- 11. Pipelines
- 12. VPC Cluster Development AZ 1 AZ 2 Cluster Integration AZ 1 AZ 2 Cluster Production AZ 1 AZ 2 Development Integration ProductionNamespaces::
- 13. Build Pipeline for docker images Bitbucket GOCD ArtifactoryUnittests ScansDEV
- 14. Deploy Pipeline Jenkins Git Kubernetes Cluster Tests Step 1 Action Pull - Branches - Jenkinsfile - Tests - Config IP based: Endpoints Logic Performance API Delete or Make public (DNS) -> Canary Tests Step 2 DNS based: Interaction Integration API UI API
- 15. Integration Tests Jenkins Git Kubernetes Cluster Tests Action Pull - Branches - Jenkinsfile - Tests - Config BV Features Performance Rollback DNS switch …and monitoring
- 16. Services
- 18. Import Analyzer Repository Search Export Storage Meta-DB Web-UI Mandanten Lieferanten Services
- 19. Import Analyzer Repository Search Export Storage Web-UI Mandanten Lieferanten <XML> <XML> Meta-DB Import Workflow - Choreographie
- 20. Import Analyzer Repository Search Export Storage Web-UI Mandanten Lieferanten <XML> <XML> Import Timeline Meta-DB Restore Edit Delete Import Workflow - Orchestration
- 22. Messages Format • JSON (Schema) Content • Message-ID • Service-Identifier • Status • Payload
- 24. Conclusion
Editor's Notes
- #3: Describe what it is, from a business perspective
- #4: USP / business features: focus the work where the need is, balance it out with efforts on operations, on testing, on deployment. These topics have to be automated, stable.
- #5: You need to be aware of this. Think of what you want to achieve in those areas. Think about the how.
- #7: Good enough: there is a common fallacy is software development ... called premature optimization