Microservices infrastructure using docker + mesos + marathon
Download as PPTX, PDF
0 likes697 views
This presentation explains about problems with conventional way of deploying microservice architecture. It also explains how do Docker + Mesos + Marathon help overcoming those problems in addition to few out of the box advantages that come along.
Microservices infrastructure using docker + mesos + marathon
- 1. Microservices Infrastructure Using Docker + Mesos + Marathon Ashwini Mutalik Desai
- 2. Monolithic vs Microservices architecture
- 3. Deploying microservices: Conventional way VPN
- 4. Problems with conventional methods Non OS specific artifacts Manual tagging/versioning and retrieving/accessing of artifacts Not straightforward to mimic an environment locally Scaling not very seamless Non-optimum resource utilization Complex fault tolerance and auto healing Log and status checking, debugging cumbersome Separate setup for monitoring
- 5. Problems with conventional methods Non OS specific artifacts Manual tagging/versioning and retrieving/accessing of artifacts Not straightforward to mimic an environment locally Build
- 6. Problems with conventional methods Non-optimum resource utilization No out of the box task level isolation Cant reschedule tasks at runtime All environments handled similarly Cluster Management
- 7. Problems with conventional methods Scaling not very seamless Complex fault tolerance and auto healing Separate setup for monitoring Log and status checking, debugging cumbersome Deployment & Orchestration
- 8. Docker + Mesos + Marathon
- 9. Docker Features OS specific images Tagged images Eases local setup Easy debugging of services (Kitematic) Enables easy mimicking of an environment locally Secure docker registry Availability of all standard packages Common interface for both dev and ops
- 10. Problems with conventional methods Non OS specific artifacts Manual tagging/versioning and retrieving/accessing of artifacts Not straightforward to mimic an environment locally Build
- 11. Mesos Features Resource isolation for tasks Efficient scheduling Running different maturity applications(dev, test, prod) Support for different kinds of workloads Web UI for monitoring Run time scheduling
- 12. Problems with conventional methods Non-optimum resource utilization No out of box task level isolation Cant reschedule tasks at runtime All environments handled similarly Cluster Management
- 13. Marathon Features Application groups and dependencies Progressive rollouts High availability and zero downtime Easy scaling Simple but powerful UI Failure handling Monitoring and healing
- 14. Problems with conventional methods Scaling not very seamless Complex fault tolerance and auto healing Separate setup for monitoring Log and status checking, debugging cumbersome Deployment & Orchestration
- 15. Build: Microservices System Using Docker Unit tests Integration tests Functional tests Contract tests Acceptance tests
- 16. Deploying microservices: Mesos + Marathon
- 17. Thank You @ashwini_md @ashwini-mutalik-desai Thank you @siddhartha91989 and @poonam-sonawane for reviewing.
Editor's Notes
- #3: Using an example explain how it can be designed in a monolithic architecture and then microservices architecture.
- #4: Rename slide? Creating and deploying a service
- #5: Animation: List all these problems first...and then in animation show them segregated by build and deploy related problems Segregate these problems at build, resource availability/performance and orchestration levels ------------- Non-optimum resource utlization - complex and involved task to make it happen in conventional way (in comparison to out of the box) http://stackoverflow.com/questions/30397050/mesos-real-world-use-cases - read resouce-utlization and preemption points
- #6: Animation: List all these problems first...and then in animation show them segregated by build and deploy related problems Segregate these problems at build, resource availability/performance and orchestration levels ------------- Non-optimum resource utlization - complex and involved task to make it happen in conventional way (in comparison to out of the box) http://stackoverflow.com/questions/30397050/mesos-real-world-use-cases - read resouce-utlization and preemption points
- #7: Animation: List all these problems first...and then in animation show them segregated by build and deploy related problems Segregate these problems at build, resource availability/performance and orchestration levels ------------- Non-optimum resource utlization - complex and involved task to make it happen in conventional way (in comparison to out of the box) http://stackoverflow.com/questions/30397050/mesos-real-world-use-cases - read resouce-utlization and preemption points
- #8: Animation: List all these problems first...and then in animation show them segregated by build and deploy related problems Segregate these problems at build, resource availability/performance and orchestration levels ------------- Non-optimum resource utlization - complex and involved task to make it happen in conventional way (in comparison to out of the box) http://stackoverflow.com/questions/30397050/mesos-real-world-use-cases - read resouce-utlization and preemption points
- #10: ----------------- Create animation for this slide and explain the following: What is docker Workflow? Why docker? Advantages Create private repository if security is the concern, pull and keep the images there What does docker registry provides build history mean? Do you want to add these? Easy on-boarding
- #11: Animation: List all these problems first...and then in animation show them segregated by build and deploy related problems Segregate these problems at build, resource availability/performance and orchestration levels ------------- Non-optimum resource utlization - complex and involved task to make it happen in conventional way (in comparison to out of the box) http://stackoverflow.com/questions/30397050/mesos-real-world-use-cases - read resouce-utlization and preemption points