introduce

  • In the interview ‘DevOpsThe father of the ‘Patrick DeboisAfter that, I deeply understood ‘DevOps is the Human Factor‘The truth of that
  • DevOpsIt’s more about practice than character
  • Pave the process through the “three-step approach”, select the right entry point, adapt the organization, continuous delivery, automation, operation and maintenance improvement according to Conway’s law
  • “Infrastructure is code” (infrastructure as code) concept
  • Operations people may become like developers, who must maintain the configuration of the system in source control and use continuous integration/continuous delivery in their work.CI/CDThe pattern of)

Common misconceptions about

  • Myth # 1:DevOpsOnly for startups. The problems they encountered were similar to those of traditional companies: software with high-risk code prone to catastrophic failures, inability to deliver new features quickly enough to defeat competitors, security compliance issues, inability to scale up services, and high levels of mistrust between development and operations
  • Erroneous zone 2:DevOpsWill replace Agile
  • Myth 3:DevOpswithITILNot compatible. Many people believe that,DevOpsReleased in 1989ITIL(Information Technology Infrastructure Library.ITInfrastructure library).DevOpsPractice can be associated withITILProcess compatibility. However, to supportDevOpsShorter release cycles and more frequent deployments,ITILMany aspects of the process need to be fully automated
  • Myth 4:DevOpsIncompatible with information security and compliance activities. Integrated into every daily part of the software development life cycle, resulting in better quality, security, and compliance
  • Myth 5:DevOpsIt means eliminationITOperation and maintenance, i.e.NoOps“. Many people make the mistake of takingDevOpsThe interpretation is complete eliminationITOperations functions, however, are rare. In this way,ITOperations people become more like developers (orQAAnd information security personnel) into the product development process, which developers use to safely and quickly test, deploy, and run in productionITService platform
  • Myth 6:DevOpsJust “infrastructure as code” or automation. butDevOpsCultural norms and structures are also needed in order toITAchieve common goals in the value stream
  • Myth 7:DevOpsThis option applies only to open source software. But to achieveDevOpsRegardless of the technology used

Introduction: OutlookDevOpsThe new world

  • Product managers, developers,QAPersonnel,ITOperation personnel and information security personnel help each other, work together, the whole company’s performance is thriving
  • The term “technical debt” isWard CunninghamFor the first time. Similar to financial debt, technical debt is when the decisions we make today cause problems that become more difficult to solve over time, with less and less action to take in the future. Even if we take on technical debt prudently, interest will still accrue
  • Development is often responsible for responding to market changes and bringing new features or changes online as quickly as possible. whileITOperation and maintenance departments to provide customers with stability, reliability and securityITServices are already in place, making it difficult or impossible for anyone to introduce changes that could harm the production environment. This configuration allows the development department andITThere is a huge conflict between the objectives and motivations of the operations department
  • The company evaluates and motivates different departments differently, which hinders the realization of the company’s overall goals
  • These chronic conflicts often result in technical workers delivering poor-quality software and services, creating poor customer experiences, and resorting to AD hoc solutions and emergencies on a daily basis

Vicious Circle trilogy

  • Part 1: Many of the problems we face in our daily work stem from applications and infrastructure that are too complex, too fragile, and not well documented. More worryingly, our most vulnerable components are supporting our most important business systems or our most critical projects
  • Step two: A product manager promises a bigger, bolder customer engagement feature, or a business executive sets a higher revenue target. This led to an increase in technical debt
  • The endgame: Everything is getting harder — everyone is busier, work is consuming more and more time, communication is slower, and work is piling up

Why are vicious cycles everywhere

  • Business leaders who want to achieve business goals are effectiveITManagement is much more dependent than they expected

Cost: people and economy

  • For our employees, this means long hours, long weekends, a diminished quality of life, and not just for them, but for all those who depend on them, including their families and friends. When this happens, it’s not surprising that we lose our best employees (except those who feel they can’t leave because of responsibility and obligation)
  • In 2011, about 5% of the worldGDP(that’s $3.1 trillion) forIT(including hardware, services and telecommunications). If we estimate that 50% of that $3.1 trillion goes to operating costs and maintaining existing systems, and one-third of that 50% goes to emergency and unscheduled work or rework

DevOpsThere is always a better way

  • By creating rapid feedback loops at each step of the process, everyone can see the results immediately
  • Automated tests can help developers find bugs quickly (usually in a matter of minutes), enable faster fixes, and real learning. If the error is discovered during integration testing six months later, the associated memory and causality have long since faded
  • Instead of days or weeks of hard work, it takes just one switch or configuration item to make a new feature work. This small change makes the new functionality visible to a larger customer base and automatically rolls back when errors occur. As a result, releasing new features becomes manageable, predictable, reversible, and less stressful
  • When something goes wrong, we conduct a post-mortem without blame, not to punish someone, but to better understand what caused the accident and how to prevent it from happening again. Because quality is important, we even deliberately inject failures into production to see how the system failed in the expected way

DevOpsBusiness value of

  • The application ofDevOpsTwenty percent of high-performing companies far underperform their peers in the following areas
  • Throughput index;
  • Number of code and change deployments (30 times frequent)
  • Code and change deployment lead time (200 times faster)
  • Reliability index
  • Production deployment (60 times higher change success rate)
  • Average service recovery time (168 times faster)
  • Organizational performance index
  • Productivity, market share, and business targets (approximately double or more)
  • Market cap growth (50% more in 3 years)
  • In other words, high performers are more agile and reliable

DevOpsHelps increase developer productivity

  • Frederick BrooksHe emphasized this point in his famous book the Myth of the Man-month. He explained that when projects are delayed, adding more developers reduces productivity not only for individual developers, but for the whole
  • On the other hand,DevOpsDemonstrates that with the right architecture, technical practices, and cultural norms in place, small development teams can quickly, safely, and independently develop, integrate, test, and deploy changes to production environments

  • The figure shows that the number of deployments per developer per day decreases in low-performing companies, stays constant in mid-performing companies, and increases linearly in high-performing companies as team size increases
  • In the applicationDevOpsAs the number of developers increases, the number of daily deployments increases linearly. Google, Amazon andNetflixAlready done
  • Another, more extreme example is Amazon. In 2011, Amazon deployed nearly 7,000 times a day; By 2015, they will be deployed 130,000 times a day

books

  • Continuous Delivery: A Systems Approach to Delivering Reliable Software
  • Goals: Common Sense Management for Simple and Effective — one of the most influential books in the lean manufacturing movement
  • Project Phoenix: OneITThe Legend of Operation and Maintenance