Why, What, and How of Continuous Delivery

Transcription

1 Why, What, and How of Continuous Delivery

2 Automation is the Key 2 Photo courtesy of Steve Jurvetson via Flickr

3 Time to Assemble a Model-T 12.5 hrs 1.5 hrs 3

4 Incredible Efficiency of Assembly Line Ford built 300,000 cars with 13,000 employees That s more than their 300 competitors combined, with 65,000 employees 4

5 5

6 Software is eating the world 6

7 7 Photo courtesy of Steve Jurvetson via Flickr

8 8 Photo courtesy of Steve Jurvetson via Flickr

9 9

10 10

11 11

12 Tomorrow is around the corner 12

13 Software is eating the world 13

14 14

15 15

16 16

17 17

18 Software is eating the world 18

19 (Millions) Evolution is required. But it is hard If you don t like change, you re going to like irrelevance even less. General Eric Shinseki Analog Digital 19

20 Nonstop demand is the #1 problem facing IT departments today. Continually optimize the IT delivery model Reorganizing IT for faster software delivery Real Time DevOps DevOps helps drive faster software development DevOps can accelerate delivery 15-20% #3. Speed of IT Delivery Agile & Incremental Software Delivery Continuous delivery 20

21 Speed + Friction = Heat 21

22 Continuous Delivery and Automation are The Answer Commit Build Test Stage Deploy Development Production Feedback Loop $ The Business 22

23 Stage Env. Discipline Continuous Delivery In Context Rapid Changes User Feedback AGILE An incremental approach to identifying, prioritizing, and coordinating feature development Continuous Integration Automated commit, build and testing of code in the development environment Continuous Delivery Software changes continuously delivered to stakeholders in any environment Continuous Deployment Software changes continuously deployed to live production Development Production / Prod-like Live Production Upstream (left) Downstream (right) Define Plan Code Compile Build Unit Test Analyze Integrate Int. Test Package Place Load Test Acct. Test Release Deploy Monitor Change Mgt. DevOps Cultural approaches & technical practices Production Bugs 23

24 Who You Wanna Be Netflix deploys a hundred times per day Amazon deploys every 11.6 seconds 24

25 Who You ACTUALLY Are Google Doc driven deployment Ops guy work through the night Deployment every once in 2 weeks 25

26 26

27 CONTINUOUS DELIVERY Nexus DEVOPS SECOPS VM LEAN CONTINUOUS DELIVERY JENKINS SCM DevOps Word Soup GIT SCRUM PERFORCE STOP THE LINE CONTINOUS INTEGRATION GIT TEST AUTOMATION SCRUM BLAZE METER XP NEW RELIC GITLAB MICROSERVICES SCRUM MICROSERVICES CONTINUOUS DEPLOYMENT CONTINUOUS IMPROVEMENT 27

28 28

29 SIMPLIFY Where are we going? Where are we? How do we get there? 29

30 30

31 Maturing Organization & Culture

32 X-Axis: SDLC Phases Agile Upstream Agile Downstream 33% 13% Define Plan Code Build Integrate Test Release Deploy Operate Source: February 5, 2014, How Can You Scale Your Agile Adoption? Forrester report 32

33 Cycle Times: Mixed Streams Agile Upstream Non-Agile Downstream 33% 87% Define Plan Code Build Integrate Test Release Deploy Operate 33

34 Cycle Times: Mixed Streams Agile Upstream Non-Agile Downstream 33% 87% Weeks, Days or Hours Define Plan Code Build Integrate Test Release Deploy Operate 34

35 Cycle Times: Mixed Streams Agile Upstream Non-Agile Downstream 33% Weeks, Days or Hours 87% Weeks to Months Define Plan Code Build Integrate Test Release Deploy Operate 35

36 Cycle Times: Mixed Streams Time-to-Market >Months Agile Upstream Non-Agile Downstream 33% Weeks, Days or Hours 87% Weeks to Months Define Plan Code Build Integrate Test Release Deploy Operate 36

37 Team Workgroup Enterprise Y-Axis: Levels of Adoption Enterprise Agile Upstream 22% Enterprise DevOps 10% Agile Upstream 33% Agile Downstream 13% Define Plan Code Build Integrate Test Release Deploy Operate Source: February 5, 2014, How Can You Scale Your Agile Adoption? Forrester report 37

38 Team Workgroup Enterprise The Destination Enterprise DevOps Innovate faster Enterprise Agile Upstream 22% Increased productivity Respond to market Gain competitive advantage Employee satisfaction and retention Agile Upstream 33% Agile Downstream 13% Define Plan Code Build Integrate Test Release Deploy Operate 38

39 The Chasms Upstream Downstream People & Culture Move Fast, Innovate Quality, Stability, Uptime Process & Practices Tools & Technology Agile, Scrum, Kanban Point tools, Grassroots, Rapid changes PMBOK, ITIL, Waterfall Enterprise class, Corp. Procurement, Stability Define Plan Code Build Integrate Test Release Deploy Operate 39

40 Team Workgroup Enterprise The Quadrants Quadrant 3: Enterprise Agile Quadrant 4: Enterprise DevOps Quadrant 1: Team-level Agile Quadrant 2: Team-level CD Define Plan Code Build Integrate Test Release Deploy Operate 40

41 The Quadrants and Adoption Patterns

42 Team Workgroup Enterprise Adoption Pattern #1 Team Agile > Ent. Agile > Ent. DevOps 1. Team(s) adopt agile planning and project management 2. Organization implements agile definition, planning, and project management 3. Organization aligns on DevOps strategy with teams using common CI/CD process and tools Define Plan Code Build/CI Test Release/Deploy Operate 42

43 Team Workgroup Enterprise Adoption Pattern #2 Agile > CD > Enterprise Agile > DevOps 1. Team(s) adopt agile planning and project management 2. Team(s) extend CI to Continuous Delivery 3. Organization implements agile 4. Organization aligns on DevOps strategy with teams using common CI/CD process and tools Define Plan Code Build/CI Test Release/Deploy Operate 43

44 Team Workgroup Enterprise Adoption Cycles: Bottom-Up Visibility Adoption Executive Awareness 2 5 Executive Support Workgroup Success 4 6 Scale Across Workgroups 1 Team Level Initiation 3 Team Successes Time 44

45 Team Workgroup Enterprise Adoption Cycles: Top-Down Visibility Adoption 1 Executive Initiative Workgroup Initiative 2 Workgroup Success 4 5 Scale Across Workgroups 3 Team Successes Time 45

46 DevOps Transformation Strategies 1. Identify pilot project 2. Form cross-functional red team 3. Adopt unifying technologies 4. Establish plan with measurable KPIs and milestones 5. Go! 6. Measure, document, refine 7. Scale and onboard 46

47 Maturing Engineering Practice

48 Infrastructure Automation Architecture Pipeline 48

49 Automation is the foundation for everything Continuous delivery (CD) is a software engineering approach in which teams produce software in short cycles, ensuring that the software can be reliably released at any time.[1] It aims at building, testing, and releasing software faster and more frequently. The approach helps reduce the cost, time, and risk of delivering changes by allowing for more incremental updates to applications in production. A straightforward and repeatable deployment process is important for continuous delivery. from Wikipedia Continuous delivery 49

50 Architecture / Implementation techniques Feature flags Activate/kill code at runtime Dark launches Put production workload without users seeing them Ease in new things Microservices Deploy different parts independently Reduce coupling 50

51 CD Pipelines that detect & reject problems dev qa production 51

52 Well-placed control points that filter out bad changes Automated validation as exit criteria from code review Reliable tests Control points that trade speed vs thoroughness in proper order 52

53 Infrastructure that enables Tolerate problems & contain damages Blue/green deployment Canary release Prevent errors Phoenix servers Immutable infrastructure 53

54 Crawl, Walk, Run 54

55 55

56 Jenkins the #1 Automation Platform for CI/CD Commit Build Test Stage Deploy Continuous Integration/Delivery Development Code Commit </> SOURCE CODE CONTROL SYSTEM Production Complex Delivery Pipelines Delivery of App and Config Robust and Highly Available Results Report Deliver Better Software Faster with CloudBees Jenkins Platform 56

57 Jenkins is the Hub of CD/DevOps Ecosystem Over 1000 Jenkins Plugins Integration with over 100 DevOps Tools Orchestration of the DevOps Toolchain End-to-End CD Pipeline Management Code & Commit Build & Config Scan & Test Release Deploy 57

58 Jenkins Popularity is Through The Roof 120, ,000 Active Jenkins Installations Worldwide Total Jenkins Jobs Worldwide 7,000,000 6,000,000 80,000 5,000,000 60,000 4,000,000 3,000,000 40,000 2,000,000 20,000 1,000, /1/ /1/ /1/ /1/ /1/ /1/ /1/ RebelLabs 2014 Java tools & technologies survey 58

59 Jenkins Installations 150,000 37% 59

60 Jenkins Workload 12,300,000 jobs 54% 60

61 # of Computers 570,000 41% 61

62 why? 62

63 It s extensible & easy to use 63

64 It connects all the things 64

65 It s a key to going fast 65

66 10 of the Fortune 100 financial services companies bank on CloudBees 66

67 7 of the Fortune 100 technology companies deliver innovation on CloudBees 67

68 2/3 of the Fortune 100 telecom companies speed up delivery with CloudBees 68

69 3 of the 4 US credit card networks speed transactions with CloudBees 69

70 Let s Eat The World And let Jenkins & CloudBees help you 70

71 THANK YOU! Software at the speed of ideas