Continuous Integration, Delivery and Deployment
4 likes809 views
The document discusses the history and evolution of continuous integration, delivery, and deployment practices in software development. It describes how the practices have developed from big bang integration, where changes were merged infrequently after long development periods, to modern approaches where integration and testing occur frequently or continuously. This allows issues to be detected and addressed earlier, reducing risks and improving productivity and code quality. Continuous delivery further aims to ensure code is always releasable, while continuous deployment automates actual releases. Surveys find these practices are becoming more common but also face challenges in adoption related to code quality, testing, organizational culture and resources.
Continuous Integration, Delivery and Deployment
- 1. Continuous Integration, Delivery and Deployment Eero Laukkanen T-76.5613 - Software Testing and Quality Assurance P 20.11.2015
- 2. System Integration In engineering, system integration is defined as the process of bringing together the component subsystems and ensuring that the subsystems function together as a system. https://en.wikipedia.org/wiki/System_integration 23/11/15 2
- 3. Software Integration • System Integration • Change Integration: Merging 23/11/15 3 Possible integration conflict!
- 4. Integration Conflicts • Merge Conflicts • Detected by version control software • Semantic Conflicts • Detected by compiler, tests and code review http://martinfowler.com/bliki/SemanticConflict.html 23/11/15 4
- 5. Regression Testing The purpose of regression testing is to ensure that changes have not introduced new faults. Experience has shown that as software is changed, emergence of new faults and/or re-emergence of old faults is quite common. https://en.wikipedia.org/wiki/Regression_testing 23/11/15 5
- 6. History of Software Integration and Regression Testing • Prehistory: Big Bang Integration • 1996: Daily Build and Smoke Test (Steve McConnell) • 2000: XP and Continuous Integration (Kent Beck) • 2006: Continuous Integration (Martin Fowler) • 2009: Continuous Deployment (Timothy Fitz) • 2010: Continuous Delivery (Jez Humble & David Farley) • 2012: Experiment System (Holmström Olsson et al.) 23/11/15 6
- 7. Big Bang Integration • System Integration • Develop subsystems independently, integrate when ready • Change Integration • Develop changes in branches, merge when ready 23/11/15 7
- 8. Big Bang Integration “We entered a huge depressing warehouse stacked full with cubes. I was told that this project had been in development for a couple of years and was currently integrating, and had been integrating for several months. My guide told me that nobody really knew how long it would take to finish integrating. From this I learned a common story of software projects: integration is a long and unpredictable process.” http://www.martinfowler.com/articles/continuousIntegration.html 23/11/15 8
- 9. Big Bang Integration • Development is simple, because nothing is changing • Integration is complex and takes unpredictable time 23/11/15 9
- 10. 23/11/15 10
- 11. 1996: Daily Build and Smoke Test • Build and smoke test the whole software system daily • Fix broken builds • “Integrate… usually once every few days” McConnell, S., 1996. Daily build and smoke test. IEEE software 143–144. 23/11/15 11
- 12. Daily Build and Smoke Test 23/11/15 12 time code developed verified Big Bang Daily • Minimizes integration and low quality risk • Easier defect diagnosis • Improved morale risk
- 13. Daily Build and Smoke Test • Context of large systems • Tests allowed to take multiple hours • Timed builds • Smoke tests evolve when the system evolves • Dedicated roles take care of the build 23/11/15 13
- 14. Daily Build and Smoke Test “The smoke test should be thorough enough that if the build passes, you can assume that it is stable enough to be tested more thoroughly.” “The standard needs to set a quality level that’s strict enough to keep showstopper defects out but lenient enough to disregard trivial defects, an undue attention to which could paralyze progress.” 23/11/15 14
- 15. 2000: Extreme Programming • Holistic software development methodology • Continuous Integration, one of the many practices Beck, K., 2000. Extreme programming explained: embrace change. Addison-Wesley Professional. 23/11/15 15
- 16. Extreme Programming and Continuous Integration “No code sits unintegrated for more than a couple of hours.” “At the end of every development episode, the code is integrated and all the tests must run at 100%.” “You need a reasonably complete test suite that runs in a few minutes.” 23/11/15 16
- 17. Continuous Integration 23/11/15 17 Develop (a couple of hours) Integration & Testing (a few minutes) Fix broken build pass fail
- 18. 2006: Continuous Integration Extended • Single source repository • Integration machine • Fix broken builds immediately • Keep the build fast (10 minutes) • Test in a clone of the production environment • Automate deployment http://www.martinfowler.com/articles/continuousIntegration.html 23/11/15 18
- 19. 2009: Continuous Deployment “The high level of our process is dead simple: Continuously integrate (commit early and often). On commit automatically run all tests. If the tests pass deploy to the cluster. If the deploy succeeds, repeat.” http://timothyfitz.com/2009/02/08/continuous-deployment/ http://timothyfitz.com/2009/02/10/continuous-deployment-at-imvu-doing-the-impossible-fifty-times-a-day/ 23/11/15 19
- 20. Continuous Deployment “So what magic happens in our test suite that allows us to skip having a manual Quality Assurance step in our deploy process? The magic is in the scope, scale and thoroughness. It’s a thousand test files and counting. 4.4 machine hours of automated tests to be exact.” “Great test coverage is not enough. Continuous Deployment requires much more than that. Continuous Deployment means running all your tests, all the time. That means tests must be reliable.” 23/11/15 20
- 21. Continuous Deployment • Context of cloud systems • Problems in production will always happen • How to mitigate the problems? • Smaller releases have smaller scope and are easier to debug • Automated deployments allow fast fixing • Deploy to a subset of users first to mitigate problem scope • Production can be monitored and reverted automatically 23/11/15 21
- 22. Continuous Deployment 23/11/15 22 developed verified released Frequent verification time code Frequent release risk • Minimizes release risk • Easier defect diagnosis • Improved morale
- 23. 2010: Continuous Delivery • Every change should be releasable… • ...but not necessarily released automatically • Allows human verification Humble, J., Farley, D., 2010. Continuous Delivery: Reliable Software Releases through Build, Test, and Deployment Automation. Addison-Wesley Professional. 23/11/15 23
- 26. Continuous Delivery You’re doing continuous delivery when: • Your software is deployable throughout its lifecycle • Your team prioritizes keeping the software deployable over working on new features • Anybody can get fast, automated feedback on the production readiness of their systems any time somebody makes a change to them • You can perform push-button deployments of any version of the software to any environment on demand http://martinfowler.com/bliki/ContinuousDelivery.html 23/11/15 26
- 27. 2012: Experiment System 23/11/15 27 Traditional Development Agile R&D Organization Continuous Integration Continuous Deployment R&D as an Experiment System Olsson, H.H., Alahyari, H., Bosch, J., 2012. Climbing the “Stairway to Heaven” – A Multiple-Case Study Exploring Barriers in the Transition from Agile Development Towards Continuous Deployment of Software, in: Proceedings of the 2012 38th Euromicro Conference on Software Engineering and Advanced Applications. Washington, DC, USA, pp. 392–399.
- 28. Experiment System “actual deployment of software functionality is seen as a way of experimenting and testing what the customer needs” • Beyond regression testing and integration • Avoid software bloat 23/11/15 28
- 31. Continuous Integration 23/11/15 31 Develop + Integrate ReleaseQA
- 32. Continuous Delivery 23/11/15 32 Develop + Integrate + QA Release
- 33. Continuous Deployment 23/11/15 33 Develop + Integrate + QA + Release
- 34. Practitioner surveys • 2015 State of DevOps Report • ~5000 respondents • 2013 Continuous Delivery: A Maturity Assessment Model • ~300 respondents https://puppetlabs.com/2015-devops-report http://info.thoughtworks.com/Continuous-Delivery-Maturity-Model.html 23/11/15 34
- 36. Survey: deployment lead time 23/11/15 36
- 39. Problems when adopting continuous delivery • Build design • Complex build, inflexible build • Product design • Modularization, internal dependencies, database schema changes, unsuitable architecture • Integration • Large commits, broken build, merge conflicts, work blockage, long-running branches… • Testing • Ambiguous test result, flaky tests, untestable code, time-consuming testing, UI testing… • Release • Customer data preservation, documentation, feature discovery, marketing, more deployed bugs • Human & Organization • Lack of discipline, more pressure, lack of motivation, lack of experience, team coordination… • Resources • Effort, insufficient hardware resources, network latencies Laukkanen, E., Itkonen, J., Lassenius, C., 2015. Problems, Causes and Solutions When Adopting Continuous Delivery - A Systematic Literature Review. Manuscript in preparation. 23/11/15 39
- 40. Problems are connected 23/11/15 40 Laukkanen, E., Itkonen, J., Lassenius, C., 2015. Problems, Causes and Solutions When Adopting Continuous Delivery - A Systematic Literature Review. Manuscript in preparation.
- 41. What to consider • Value of fast release cycle • Competitive market • Product ambiguity • Cost of fast release cycle • Competences, legacy code, infrastructure, maintenance • How often can you release • Cost of release • How long can quality assurance take • Cost of production bugs 23/11/15 41