DevOps Indonesia #8 - Service Mesh with Istio
•
1 j'aime•298 vues
Slide from DevOps Indonesia #8 Meetup - Service Mesh with Istio @BTPN Jenius Office
Recommandé
Recommandé
Contenu connexe
Tendances
Tendances (20)
Similaire à DevOps Indonesia #8 - Service Mesh with Istio
Similaire à DevOps Indonesia #8 - Service Mesh with Istio (20)
Plus de DevOps Indonesia
Plus de DevOps Indonesia (20)
Dernier
Dernier (20)
DevOps Indonesia #8 - Service Mesh with Istio
- 1. PAGE1 DEVOPS INDONESIA Solutions Architect, Indonesia, Red Hat Okky Hendriansyah Tri Firgantoro Service Mesh with Istio
- 2. PAGE2 DEVOPS INDONESIA Agenda • Background • Architecture • Features • Simple Demo • Q & A
- 7. PAGE7 DEVOPS INDONESIA Microservices Application =~ Distributed Computing
- 8. PAGE8 DEVOPS INDONESIA Fallacies of Distributed Computing 1. The network is reliable 2. Latency is zero 3. Bandwidth is infinite 4. The network is secure 5. Topology doesn’t change 6. There is one administrator 7. Transport cost is zero 8. The network is homogeneous source: https://en.wikipedia.org/wiki/Fallacies_of_distributed_computing
- 9. PAGE9 DEVOPS INDONESIA What happens if a microservice failed?
- 10. PAGE10 DEVOPS INDONESIA What happens if a microservice failed?
- 12. PAGE12 DEVOPS INDONESIA Why Not Traditional Enterprise Service Bus? source: https://www.ains.com/enterprise-service-bus-esb/
- 13. PAGE13 DEVOPS INDONESIA The Solution? • A service mesh platform originally developed by Google, IBM, and Lyft on May 2017 • Built using battle-tested Envoy Proxy from Lyft • Just turned v1.0 this August 2018! • Deploys on Kubernetes and Consul- based deployment
- 14. PAGE14 DEVOPS INDONESIA What is a Service Mesh? A service mesh is a dedicated infrastructure layer for handling service-to-service communication. It’s responsible for the reliable delivery of requests through the complex topology of services that comprise a modern, cloud native application. In practice, the service mesh is typically implemented as an array of lightweight network proxies that are deployed alongside application code, without the application needing to be aware. source: https://blog.buoyant.io/2017/04/25/whats-a-service-mesh-and-why-do-i-need-one/
- 15. PAGE15 DEVOPS INDONESIA What is a Service Mesh? A configurable infrastructure layer for a microservices application. It makes communication between service instances flexible, reliable, and fast. The mesh provides service discovery, load balancing, encryption, authentication and authorization, support for the circuit breaker pattern, and other capabilities. source: https://www.nginx.com/blog/what-is-a-service-mesh/
- 16. PAGE16 DEVOPS INDONESIA Istio Approach of Service Mesh • Deploys EnvoyProxy as the default Istio Proxy as a sidecar container for each service deployed • The Istio Proxy sidecars are injected (automatically or manually) • No more single point of failure, since the proxies are distributed and have cache mechanism • Istio Proxy obtain rules and configuration from Istio Pilot
- 17. PAGE17 DEVOPS INDONESIA Why Istio Choose EnvoyProxy? • Originally developed by Lyft • High performance distributed C++ based proxy • Understands HTTP, HTTP/2, Kafka, TCP, gRPC, etc. • Robust API • Open source and donated to CNCF
- 18. PAGE18 DEVOPS INDONESIA Istio Architecture ● Control Plane ○ Pilot ○ Mixer ○ Citadel ● Data Plane ○ Proxy ● Gateway ○ Ingress ○ Egress ● Addons ○ Prometheus ○ Jaeger ○ ServiceGraph ○ Grafana source: https://istio.io
- 19. PAGE19 DEVOPS INDONESIA Pilot & Proxy Relationship ● Pilot is responsible for the lifecycle of Proxy ● Rules are defined in the Pilot ● Pilot hides the platform complexity from Proxy ● Service discovery and rules are pushed and cached from Pilot to Proxy source: https://istio.io
- 20. PAGE20 DEVOPS INDONESIA Mixer & Proxy Relationship ● Mixer provides policy control ● Mixer collects telemetry source: https://istio.io
- 21. PAGE21 DEVOPS INDONESIA Mixer & Proxy Relationship ● Proxy calls Mixer to check policy before each request ● Proxy calls Mixer to report telemetry after each request ● Proxy and Mixer cached information to reduce latency source: https://istio.io
- 22. PAGE22 DEVOPS INDONESIA Traffic Management: Intelligent Routing Content-based Routing
- 23. PAGE23 DEVOPS INDONESIA Traffic Management: Traffic Shifting Canary Deployment, Blue/Green Deployment, Weight-based Routing
- 24. PAGE24 DEVOPS INDONESIA Traffic Management: Ingress & Egress Traffic Ingress, Service-to-Service, and Egress By default, outgoing connection outside cluster is denied
- 25. PAGE25 DEVOPS INDONESIA Traffic Management: Split Traffic Between Versions Proxy will forward to specific version based on routing rules
- 26. PAGE26 DEVOPS INDONESIA Other Features (not covered) • Fault Injection • Circuit Breaker • Delays/Timeout • Retries • Security (encryption) • Policy (rate limit, whitelist, blacklist, denial) • Telemetry • etc.
- 27. PAGE27 DEVOPS INDONESIA Simple Demo: Bookinfo
- 28. PAGE28 DEVOPS INDONESIA Simple Demo: Bookinfo
- 29. PAGE29 DEVOPS INDONESIA Alone We are smart, together We are brilliant THANK YOU ! Quote by Steve Anderson