The Rationale for Continuous Delivery
summary
Argues continuous delivery by showing concrete results: a bank with a 103‑day change cycle vs an exchange with a 57‑minute cycle, Amazon deploying on average every 11.6 seconds, Google running ~60 million builds/year and 100M tests/day. Explains a deployment pipeline: commit-triggered fast CI (tests under ~5 minutes), store built artifacts, automated acceptance and performance tests in production‑like environments, kill failing release candidates, and make continuous deployment a business decision.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:48The State of Software Development Has Been Err.... Sub-Optimal
- 2:20What Have We Tried?
- 3:07Learning From Our Mistakes
- 4:10What Do We Really Want?
- 4:42A Question....
- 5:22The Scientific Method
- 7:30What Works? - More Data
- 10:01What Really Works?
- 13:10Cycle-Time
- 16:15What Is Continuous Delivery?
- 19:42The Principles of Continuous Delivery
- 26:43What Does This Look Like?
- 33:17Example Continuous Delivery Process
- 42:57HP LaserJet Firmware Team
- 45:02The Results
- 45:36The Effect on Business - Part 1
- 46:55Who Practices CD?
related talks
33mContinuous Delivery - What does good really look like?
24mWhy, What, and How of Continuous Delivery
40mSustainable Continuous Delivery: Patterns for rapid and reliable Releases
51mBetter software, faster: principles of Continuous Delivery and DevOps
59mThe continuous culture
36mWhat do we know about continuous delivery – and what would we like to know
47mIt's Not Continuous Delivery if You Can't Deploy Right Now
28mMinimum Viable Continuous Delivery