How do I do DevOps when all I have is Ops?
0 likes4,909 views
The document discusses how to implement DevOps practices when an organization only has operational capabilities. It recommends focusing on operational data mining to enable continuous learning and experimentation. Specific approaches include training operations staff on infrastructure as code using tools like Git, Ansible, and Jenkins. Emphasis is placed on empowering frontline staff by giving them insights from operational data to identify constraints and make improvements. The overall goal is to help organizations gain the agility benefits of DevOps through incremental cultural and process changes rather than waiting for full development capabilities.
How do I do DevOps when all I have is Ops?
- 1. © 2016 Computer Sciences Corporation How do I do DevOps when all I have is Ops? Chris Swan @cpswan CTO Global Infrastructure Services
- 2. 2 October 14, 2016© 2016 Computer Sciences Corporation
- 3. 3 October 14, 2016© 2016 Computer Sciences Corporation Agenda • Organisational Context • What do I mean by DevOps? • All in on Operational Data Mining • People, Process and Tools
- 4. © 2016 Computer Sciences Corporation The organisational context
- 5. 5 October 14, 2016© 2016 Computer Sciences Corporation Let’s get Conway’s law out of the way early Dev Ops GBS GIS
- 6. 6 October 14, 2016© 2016 Computer Sciences Corporation iDo – how we presently organise delivery
- 7. © 2016 Computer Sciences Corporation What do I mean by DevOps?
- 8. 8 October 14, 2016© 2016 Computer Sciences Corporation All industries evolve Design for Purpose Design for Operations Design for Manufacture
- 9. 9 October 14, 2016© 2016 Computer Sciences Corporation IT is no different Design for Purpose Design for Operations Design for Manufacture
- 10. 10 October 14, 2016© 2016 Computer Sciences Corporation Agility, Agility, Agility High-performing IT organizations report experiencing: 200x 2,555x more frequent deployments lower change failure rate shorter lead times faster recovery from failures 24x 3x Concave Convex Gains tail off at scale Typical of most traditional systems Gains grow at scale Typical of most software based systems Payoff Success Payoff Success Source: 2016 State of DevOps Report Business Agility Agile Software Platform Agile Infrastructure
- 11. 11 October 14, 2016© 2016 Computer Sciences Corporation The 3 ways 1.Flow 2.Feedback 3.Continual Learning & Experimentation
- 12. 12 October 14, 2016© 2016 Computer Sciences Corporation Flow – people want Continuous Delivery, and we can help with that, but it’s not something that can be assumed or directly influenced You Are Here
- 13. 13 October 14, 2016© 2016 Computer Sciences Corporation Feedback – it’s pretty much the same story as with Flow
- 14. 14 October 14, 2016© 2016 Computer Sciences Corporation Operational Data Mining (ODM) provides a means for Continual Learning and Experimentation ‘Exhausting’ by Ben Salter https://flic.kr/p/8VTaMe
- 15. © 2016 Computer Sciences Corporation All in on ODM
- 16. 16 October 14, 2016© 2016 Computer Sciences Corporation Data helps us find the constraints, then tells us what to do with them ‘Narrow’ by gwire https://flic.kr/p/4d3N4
- 17. 17 October 14, 2016© 2016 Computer Sciences Corporation Data provides a means of empowerment to front line staff “I knew that, I knew that we needed to do that”
- 18. © 2016 Computer Sciences Corporation People, Process and Tools
- 19. 19 October 14, 2016© 2016 Computer Sciences Corporation Infrastructure as Code boot camp
- 20. 20 October 14, 2016© 2016 Computer Sciences Corporation Infrastructure as Code boot camp - Git
- 21. 21 October 14, 2016© 2016 Computer Sciences Corporation We pay particular attention to Fork’n’Pull Beg forgiveness Ask permission
- 22. 22 October 14, 2016© 2016 Computer Sciences Corporation Infrastructure as Code boot camp - Ansible
- 23. 23 October 14, 2016© 2016 Computer Sciences Corporation Infrastructure as Code boot camp - Jenkins
- 24. 24 October 14, 2016© 2016 Computer Sciences Corporation The perfect runbook… ‘Snug’ by Thomas https://flic.kr/p/qbdT3c
- 25. 25 October 14, 2016© 2016 Computer Sciences Corporation Doing root cause analysis properly Why?
- 26. © 2016 Computer Sciences Corporation Wrapping up
- 27. 27 October 14, 2016© 2016 Computer Sciences Corporation Review • Organisational Context • What do I mean by DevOps? • All in on Operational Data Mining • People, Process and Tools
- 28. 28 October 14, 2016© 2016 Computer Sciences Corporation Further reading
- 29. 29 October 14, 2016© 2016 Computer Sciences Corporation Final Thought
- 30. © 2016 Computer Sciences Corporation Questions?