µServices Architecture @ EPAM WOW 2015
6 likes1,272 views
The document discusses the principles and implementation of microservices architecture, emphasizing its benefits such as speed, availability, and scalability. It covers best practices in design, deployment, and operation, as well as tools like Dropwizard, Spring Boot, and Vert.x that facilitate this architecture. Additionally, it highlights the need for a culture of trust and responsibility among developers, while acknowledging the complexities of polyglot infrastructure and operational responsibilities.
µServices Architecture @ EPAM WOW 2015
- 1. μServices Architecture @ EPAM WOW 2015
- 2. ● SA at EPAM Systems ● primary skill is Java ● hands-on-coding with Groovy, Ruby ● trying to get in touch with Erlang/Elixir ● passionate about agile, clean code and devops Izzet Mustafayev@EPAM Systems @webdizz webdizz izzetmustafaiev http://webdizz.name
- 3. Agenda ● What is this? ● Architecture ● Case study ● Existing tools ● Summary ● How to get there? ● Q&A
- 6. ● Small (100s LOC)* Overview
- 7. ● Small (100s LOC)* ● Lightweight (run several per box) Overview
- 8. ● Small (100s LOC)* ● Lightweight (run several per box) ● Takes strength of platform/language (polyglot) Overview
- 9. ● Small (100s LOC)* ● Lightweight (run several per box) ● Takes strength of platform/language (polyglot) ● Independent (development/deployment) Overview
- 10. ● Small (100s LOC)* ● Lightweight (run several per box) ● Takes strength of platform/language (polyglot) ● Independent (development/deployment) ● Stateless (everything persisted/not persisted) Overview
- 11. ● Small (100s LOC)* ● Lightweight (run several per box) ● Takes strength of platform/language (polyglot) ● Independent (development/deployment) ● Stateless (everything persisted/not persisted) ● Monitored (health and business value) Overview
- 12. ● Design ● Implementation ● Delivery ● Operation Architecture
- 13. Design ● What ab SOA? ● Domain-Driven Design ● Versioning ● Integration ● Reusability ● Scalability ● Availability ● Asynchronicity ● Externalization of Configuration
- 14. What about SOA? ● Sensible idea ● Lack of guidance ● Combat challenges of large monolithic applications
- 15. Domain-Driven Design Set of principles and patterns that ● are closed to business ● are closed to developers ● to craft elegant object systems
- 16. Versioning ● Leave multiple old microservice versions running ● Fast introduction vs. slow retirement asymmetry
- 17. Integration ● Own service - own client ● Client load balancer ● Thin and efficient transport/protocol
- 18. Reusability ● Cohesive logic ● Usage of Components/Frameworks
- 20. ● Fault-tolerance ● Fast and small failures Availability
- 21. Asynchronicity ● Non-blocking operations ● Reactive principles ● Messaging
- 22. Externalization of Configuration ● Support for X > 1 environments ● Run-time changes
- 23. ● Separation of Concerns ● Caching Implementation
- 24. Separation Concerns ● Inverse Conway’s law – teams own service groups
- 25. Separate Concerns ● Inverse Conway’s law – teams own service groups ● Each service is it’s own build
- 26. Separate Concerns ● Inverse Conway’s law – teams own service groups ● Each service is it’s own build ● Lightweight mocked dependencies
- 27. Separate Concerns ● Inverse Conway’s law – teams own service groups ● Each service is it’s own build ● Lightweight mocked dependencies ● Easy to run locally
- 28. Caching ● Stateful cached data access layer
- 29. Caching ● Stateful cached data access layer ● Expensive computational data
- 30. Caching ● Stateful cached data access layer ● Expensive computational data ● Static resources/CDN/Mirroring
- 31. ● Immutability ● Disposability ● Containerisation ● Zero-down-time ● Frequency Delivery
- 32. ● Build ● Infrastructure ● Infrastructure-as-code Immutability
- 33. ● Build ● Run ● Destroy Disposability
- 34. ● Container as deployment artifact ● Environment agnostic ● New version - new container ● All dependencies built in Containerisation
- 35. ● Blue-green ● Multi-phased ● Feature toggle ● Database migration/versioning Zero-down-time
- 36. Frequency ● Fast rollouts ● Fast rollbacks
- 37. ● Service Discovery ● Technical Monitoring ● Business Monitoring ● Stats Operation
- 38. ● Services use services ● Each service has own version ● Decoupling still matters ● One-the-fly parameterisation Service discovery
- 39. ● Hardware monitoring ● Resources utilisation ● Health checking Technical monitoring
- 40. ● Business aware metrics ● A/B testing metrics Business monitoring
- 41. ● Increased traffic velocity makes individual events less important ● More vertical components require more measurement ● Failure Rate is more informative than individual failures Stats matter on scale
- 42. Benefits
- 43. ● Polyglot technology stack ● Polyglot persistence ● Frameworks ● Thin transport Toolset unchained
- 44. ● Independent provisioning ● Fine tuning ● Elasticity Scalability
- 45. ● Development ● Testing ● Deployment ● Reliability Independence
- 46. Shortcomings
- 47. ● More responsibility from Devs to support Ops Shortcomings
- 48. ● More responsibility from Devs to support Ops ● Polyglot infrastructure (if any) Shortcomings
- 49. ● More responsibility from Devs to support Ops ● Polyglot infrastructure (if any) ● Orchestration Shortcomings
- 50. ● More responsibility from Devs to support Ops ● Polyglot infrastructure (if any) ● Orchestration ● Costs Shortcomings
- 51. ● More responsibility from Devs to support Ops ● Polyglot infrastructure (if any) ● Orchestration ● Costs ● Dependency management Shortcomings
- 52. ● More responsibility from Devs to support Ops ● Polyglot infrastructure (if any) ● Orchestration ● Costs ● Dependency management ● Distributed applications are complicated Shortcomings
- 54. Twitter before
- 56. Twitter now*
- 57. Twitter requests
- 58. ● Speed wins in the marketplace Lessons learned
- 59. ● Speed wins in the marketplace ● Remove friction from product development Lessons learned
- 60. ● Speed wins in the marketplace ● Remove friction from product development ● High trust, low process Lessons learned
- 61. ● Speed wins in the marketplace ● Remove friction from product development ● High trust, low process ● Freedom and responsibility culture Lessons learned
- 62. ● Speed wins in the marketplace ● Remove friction from product development ● High trust, low process ● Freedom and responsibility culture ● Simple patterns automated by tooling Lessons learned
- 63. ● Speed wins in the marketplace ● Remove friction from product development ● High trust, low process ● Freedom and responsibility culture ● Simple patterns automated by tooling ● Microservices for speed and availability Lessons learned
- 64. ● Speed wins in the marketplace ● Remove friction from product development ● High trust, low process ● Freedom and responsibility culture ● Simple patterns automated by tooling ● Microservices for speed and availability ● Use statistics to monitor behavior Lessons learned
- 65. Toolset ● Dropwizard ● Spring Boot ● Vert.X ● More...
- 66. Dropwizard http://dropwizard.github.io/dropwizard/ Java framework for developing ops- friendly, high-performance, RESTful web services. - pulls together stable, mature libraries from the Java ecosystem into a simple, light-weight package - has out-of-the-box support for sophisticated configuration, application metrics, logging, operational tools, and much more
- 67. Spring Boot http://projects.spring.io/spring-boot/ Takes an opinionated view of building production-ready Spring applications. - favors convention over configuration and is designed to get you up and running as quickly as possible. - production-ready features such as metrics, health checks and externalized configuration
- 68. Vert.X http://vertx.io/ ● Lightweight, reactive, application platform ● Superficially similar to Node.js - but not a clone! ● Inspired also from Erlang/OTP ● Polyglot ● High performance ● Simple but not simplistic
- 69. More... ● Ratpack http://www.ratpack.io/ ● Sinatra http://www.sinatrarb.com/ ● Spark http://sparkjava.com/ ● Compojure https://github.com/weavejester/compojure ● Webbit https://github.com/webbit/webbit ● Finagle http://twitter.github.io/finagle/ ● Connect http://www.senchalabs.org/connect/ ● XDropWizard https://github.com/timmolter/XDropWizard
- 70. Summary
- 71. Summary ● Speed wins in the marketplace
- 72. Summary ● Speed wins in the marketplace ● Separated concerns at app level
- 73. Summary ● Speed wins in the marketplace ● Separated concerns at app level ● Focused and automated
- 74. Summary ● Speed wins in the marketplace ● Separated concerns at app level ● Focused and automated ● Elastically scalable
- 75. Summary ● Speed wins in the marketplace ● Separated concerns at app level ● Focused and automated ● Elastically scalable ● Polyglot infrastructure
- 76. ● Automate repeating activities ● “Run what you wrote” – root access and duty ● Freedom and responsibility for developers How to get there?
- 77. References ● http://martinfowler.com/articles/microservices.html ● http://yobriefca.se/blog/2013/04/28/micro-service- architecture/ ● http://microservices.io/ ● http://goo.gl/TmZ4Ee ● http://12factor.net/ ● http://goo.gl/tnPxLP
- 78. Q&A
- 80. Izzet Mustafayev@EPAM Systems @webdizz webdizz izzetmustafaiev http://webdizz.name