Open Microservice Architecture

The microservice architecture approach has been very popular in the recent years. There is a big hype around it and a large swarm of open source tools to facilitate each aspect of this architecture. The purpose of this talk is to identify the main components of a microservice architecture. After that we compare different open source tools that fits into each area. At the end we’ll have a good understanding what a microservice architecture based on OSS looks like.

Transcript

1. Open Microservice Architecture @stoitsev
3. API
4. Components Cities Users Payments Products Trips Receipts
5. Services Cities Users Payments Products Trips Receipts
6. “Loosely coupled service oriented architecture with bounded contexts” Adrian Cockcroft
7. Users Products Trips Cities Payments Receipts
8. Deployment Communication Monitoring
9. Containers
11. cgroups
12. namespaces
13. https://github.com/docker/docker
14. Docker Swarm https://github.com/docker/swarm
15. Kubernetes https://github.com/kubernetes/kubernetes
16. Node Node Pod Container Container Container Container Container Container Container
17. Mesos http://mesos.apache.org/
18. http://mesos.apache.org/documentation/latest/architecture/
19. Mesos at Uber
20. RPC → DCOM, CORBA → RMI → XML-RPC → SOAP → REST → REST, Thrift
21. Thrift https://thrift.apache.org/
22. Discovery
23. Etcd https://github.com/coreos/etcd
24. etcd etcd
25. Hyperbahn https://github.com/uber/hyperbahn
26. Load balancing
27. Service Pod Pod Pod Pod Users Users Users Users
28. Ringpop https://github.com/uber/ringpop-node
30. Circuit breaking
31. Monitoring
32. What to monitor? • Success rate • Error rate • Response time - p95, p99 • RPS • Queue size • Memory usage • Disk space usage • ...
33. Business metrics!!!
34. Kafka hadoop ELK
35. statsd https://github.com/etsy/statsd https://github.com/uber/lynx
36. Monitoring is not enough
38. Illustration © Lev Polyakov, http://levpolyakov.com
39. Thanks!