Microservices, Containers, and Beyond
2 likes174 views
Microservices, containers, and serverless technologies can help build scalable distributed systems. Microservices break applications into independently deployable components organized around business capabilities. Containers provide isolation and repeatability. Kubernetes orchestrates groups of containers into pods and services. Service meshes like Istio help govern microservices through sidecar proxies. Serverless platforms automate resource management so code runs without servers. Knative in particular helps scale microservices to zero or many instances.
Microservices, Containers, and Beyond
- 1. Microservices, Containers, and Beyond Senior Director Cloud Architecture, WSO2 @lakwarus Lakmal Warusawithana
- 2. Microservice Architecture Image source : https://martinfowler.com/articles/microservices.html
- 3. Microservice Characteristics Componentization as a Service ● One main reason for using services as components (rather than libraries) is that services are independently deployable Organized Around Business Capabilities ● Segregating capabilities for specific business areas like user interface and external integrations Smart Endpoints and Dumb Pipes ● Each microservice is as decoupled as possible with its own domain logic
- 4. Microservice Characteristics Image source : https://martinfowler.com/articles/microservices.html
- 5. Microservice Characteristics ● Not influencing any programing languages, data types etc ● Use whatever works for you ● But in practice, using a cloud native programing language is more productive
- 6. Microservice Characteristics “Micro” in microservice isn’t about the no. of lines of code ● The problem should be small enough to be handled by a team ● Team should be fitting into two pizzas ● No tool can help find the right scope
- 7. Lessons learnt from SOA ● Don’t do it because everyone else is doing ● Look for a real use case Use the correct strategies to move monolithic to microservice Mindful
- 8. Mindful Need to have a proper balance ● Don’t over complicate ● Agility and repeatability is the key ● Choose the correct tools and framework whenever possible
- 9. Microservice Challengers We are building a distributed system ● Complex service discovery ● Network resiliency ● Failover and load balancing ● Global testing ● Global monitoring ○ Global tracing ○ Global logging ● Governance and policy enforcement ● Patching, updates and new versions rolling out
- 11. Containers (Docker) Provide isolation, but more controllability and repeatability Universal packaging model with dependencies and configurations Containers: ● Lightweight ● Fast Boot Up ● Immutable Single process in single container
- 12. Reality However in many cases, a single container is not enough to run an application
- 14. Small group of containers and volume Tightly coupled ● The atom of replication and placement Logical host for containers ● Echo pod get an IP address ● Share data: localhost, volume, IPC, etc. K8S POD
- 15. Group of pods that work together DNS name Load balanced: name map to a stable virtual IP Decoupled from Pods and Replication controllers K8S Service
- 17. Blue / Green Image source: https://www.gocd.org/2017/07/25/blue-green-deployments.html
- 18. Canary Image source : https://www.gocd.org/2017/08/15/canary-releases/
- 20. Service Mesh Image source : http://philcalcado.com/2017/08/03/pattern_service_mesh.html Is mesh a Distributed ESB? Trying to solve ● Governance and Policy enforcement ● Network resiliency ● Service discovery ● Version control ● Deployment and update orchestration
- 21. Sidecar Proxy Run adjacent to the main service ● Running in the same Pod ● But as separate container Capable of handling ● Ingress ● Egress ● Service discovery ● Circuit breaker ● Failover, load balancing
- 22. Sidecar Proxy Image source: http://philcalcado.com/2017/08/03/pattern_service_mesh.html
- 25. Serverless What is Serverless? ● No need to have servers to run our apps Where are our apps running?
- 26. Serverless There are still servers (containers) ● But, they are not running always ● You don’t manage them anymore Because of that ● No need to optimize them ● No need to worry about HA and scaling
- 27. Serverless vendors and technologies ● Amazon Lamba ● Azure Function ● Google Function ● IMB Functions ● Serverles.com ● Openwiks ● Kubeless ● Knative
- 29. Demo: Knative - zero to N scaling