Summary of DevOps Transformation Experiences of Large Technological Organizations

 The basic conditions for DevOps Any practice has its birth and application of soil. As with the continuous integration of agile central practices, DevOps is the best practice for cloud platform development and delivery. In other words, it is because of two major changes brought by the cloud platform: 1. Simplifies the operation and maintenance of IT infrastructure, but increases the size of the server under management. 2. The consistency of each environment (network, operating system version, the previous production environment License restrictions, the difficulty of preparing the same environment, you can now use image / snapshot solution), allows developers to use a small amount of wescript to complete the deployment of different environments, Even control the deployment and recovery of resources. Two major changes have led to a reduction in non-cross-cutting areas of expertise in development and operation. This has the potential to unify the responsibilities around the ultimate business goal: the right software. But the most fundamental reason for this is the explosive growth of Internet applications. Instagram, for example, added 10 million users in 10 days and added 60 million users in 5 months. Even in the cloud architecture (Instagram was initially set up on AWS), there is no set of advanced software engineering practices, so many users of photo uploads, compressed storage and access, means that the O & M engineering team must rapidly expand, and traditional hiring And the induction process, it is difficult to have any effect within ten days, in this case, the business can only look disappointed! Large technical organizations A brief sentence: is too much. In the past, when the scale of business was increased, technical organizations of more than a thousand people did not have the overall quality level after scale-up
guaranteed by simplification or governance at several levels like the good example of Instagram above:  Is the infrastructure scalable?  Is the architecture scalable?  Are core competencies shared in the organization?  Avoid manual errors? Of course, lack of a clear business value proposition is one of the reasons why technology organizations are getting larger, like Instagram, which does not become a behemoth when its users grow by 70 million and is still a photo-sharing application, linked to its clear value proposition , But this topic is not discussed in this article. Therefore, in the process of becoming larger in technological organization, the problems usually faced are:  Old technology, a lot of debt  Staff capacity is uneven, the team often laborless but fruitless  Online quality accidents, business satisfaction is low Practice can not drive organizations As mentioned earlier, DevOps is a software engineering practice that for today is one of the most advanced software engineering practices, implying the ability to consistently deliver reliable business versions at low cost and short cycles. Organization is not a project. Practice can change one's cognition of the process of achieving results. However, organization is a link that goes beyond the individual, and practice can not make any change to the organization. For the organization, DevOps implementation is a comprehensive challenge. Let's look at the DevOps coverage given by the wiki: 1. Code - code development and review, source code management tools, code merging 2. Build - continuous integration tools, build status 3. Test - continuous testing tools that provide feedback on business risks 4. Package - artifact repository, application pre-deployment staging 5. Release - change management, release approvals, release automation 6. Configure - infrastructure configuration and management, Infrastructure as Code tools 7. Monitor - applications performance monitoring, end-user experience 8. It requires a consistent set of deployment standards and tools for a consistent deployment environment; it requires teams to be disciplined in their code integration efforts, to write automated test capabilities, and to fix problems early. And, automate everything. 9. This means continuously challenging existing standards, processes, tools, and even security policies in the organization (opening up various deployment environments); it also means continually challenging the team's legacy of project achievements, working methods, and even experiences that once prided itself Curriculum vitae. The first layer of resistance: the lack of consensus on the root causes of the problem Many organizations have fallen by the wayside of automated testing during agile transformations. Lies in the lack of analysis and understanding of fundamental problems and underlying causes (Systemic root cause analysis). Why do automated testing? Is the current deliverable quality, especially the current version of the quality delivered to the testers?