Shulou(Shulou.com)06/02 Report--

DevOps is one of the hottest topics in the field of IT. Understanding, mastering and applying DevOps is of great significance to improve software delivery and control. However, up to now, there is no unified definition of DevOps.

This paper mainly analyzes and expounds the birth background, definition evolution and landing value of DevOps from the development of business and IT, in order to discuss the connotation of DevOps.

It has been ten years since DevOps was proposed in 2009. In recent years, the popularity of DevOps has increased rapidly, and it can also be seen from the annual report of DORA that by the end of 2018, the application of DevOps in the world's major industries has reached 30%, rising rapidly.

However, up to now, there is still no accurate definition of DevOps, which will bring great trouble to our understanding of DevOps, or in the process of landing DevOps, and all parties have their own understanding. So, what exactly is the connotation of DevOps? We look forward to a discussion from the development background, definition evolution and landing value of DevOps.

Background of DevOps development

Dependence of business development on software

The application of civil software system rose in the eighties of last century and developed vigorously after 2000. among them, as an important part of software system, it has gradually penetrated into every corner of society. From the overall development, the relationship between business and software system can be divided into three stages.

Mild dependence

In this stage, it mainly appears in the early stage, the software system mainly solves the problem of too much repetition or large amount of calculation in the business, and the scope of support is mainly limited to a certain process or link in the business. In other words, at this stage, the business can be carried out without the software system, the software is not a necessary option for the business, and the software functions can be updated frequently for days or even months or years.

Heavy dependence

With the development of the business and the increase of the relevance and complexity of the business, it has become impossible for the business to rely entirely on manual completion. To this period, the software system has become the core support of the business, the development of the business has been inseparable from the software system, but it can still accept the existence of a short non-service period, and the update frequency of software functions is required in a few days and months.

Completely dependent

As the software service goes deep into every corner of the society, the clothing, food, housing, transportation and use of the social life all depend on the software system. From a certain field or application, the software system has entered the dominant stage relative to the business. The software system must provide 365-724 services. Any interruption of service may bring great economic or social losses, and the update frequency of software functions must be controlled in a few days, hours and minutes.

IT Management and the Evolution of R & D Mode

The evolution of IT management and R & D model is closely related to the response efficiency of IT to business, which can be divided into three stages.

Steady-state mode

In the management of the traditional steady-state mode, the business, R & D, testing, and operation and maintenance are managed in stages. The personnel of each stage only need to pay attention to the relevant contents within this stage, and the R & D aims at the delivery of the system software package. The overall structure is shown in the following figure:

In the R & D model, the waterfall model is the main feature, and its advantage is that the division of each stage is relatively clear, the overall cost is lower, but the iterative speed is slow, and its applicability is better in the stage of slight dependence on software.

"steady state + agile" mode

With the increase of business dependence on software, the iterative efficiency of software in the traditional mode has become a major problem that hinders business development, so a series of optimization measures have emerged for the most time-consuming software development process. Agile R & D is its representative. During this period, the IT management model has become a "steady + agile" model, and research and development is still aimed at system package delivery, as shown in the following figure:

Of course, during this period, the efficiency improvement for each stage is also in full swing, and a large number of tools and software for different stages have emerged, as shown in the following figure:

"sensitive" mode

When the business is completely dependent on the software system, the IT management and R & D mode needs to enter the "sensitive state" mode. And this model is the DevOps that we often talk about today. In this model, the goal of R & D delivery is no longer system software packages, but business-oriented service capabilities, as shown in the following figure:

As the relationship between business and software is getting closer and closer, the delivery of research and development has shifted from software package products to service capabilities for business needs. however, this service capability also involves many aspects. let's take a look at the content-rich process of DevOps from the definition and development of DevOps.

Evolution of DevOps definition

The "steady state + agile" model can only improve the local efficiency, while the only way for DevOps is to optimize and improve the whole research and operation process as an organic whole. Since dependence was put forward in 2009, the connotation of DevOps is also constantly developing and enriching. Let's first take a look at the definition of different periods.

In October 2009, [Patrick Debois] DevOps (combination of Development and Operations) is a group of processes, methods and systems that facilitate communication, collaboration and integration between development (application / software engineering), technology operations and quality assurance (QA) departments.

In May 2010, DevOps is an important theory. Through agile, lean and other methods, in a traditional organization, independent departments work together to achieve cross-departmental integration and establish closer multi-departmental cooperation. DevOps is not only concerned with the technical issue of software deployment, it is a management method that considers communication and collaboration between departments.

December 2014 DevOps is a software development approach that emphasizes communication, collaboration and integration between software developers and information technology personnel.

November 2015 DevOps is a culture, movement, or practice that emphasizes collaboration and communication between software developers and other information technology personnel, as well as the process of automating software delivery and infrastructure changes.

AWS DevOps definition in 2018: DevOps is a set of practices designed to quickly import changes into a production environment while ensuring high quality.

We organize the definitions of different periods into a timeline and do some keyword extraction:

It can be seen that with the passage of time, the connotation of DevOps has been continuously enriched, and it has developed from the simple definition of the combination of processes, methods and systems to a system with theory, method, culture and practice.

The DevOps system is essentially different from the traditional model, which is mainly reflected in the following aspects:

Different ways of organization

DevOps emphasizes the integrated management of projects or products as the basic management mode, so as to eliminate the phased task management and execution in the traditional mode, and eliminate the convergence of different stages in the traditional mode through the adjustment of the management mode.

Focus on different points

DevOps aims at the ultimate delivery of service capabilities to the business, rather than phased task delivery in the traditional model.

Promotion dimensions are different.

DevOps pays attention to the whole life cycle of the software system, which is to improve the quality and efficiency of software delivery as a whole, not only the improvement of a certain stage, but also the promotion of each stage can be more organically coordinated.

Different depths of management and control

DevOps emphasizes full-process monitoring, full-process measurement, and transparent software delivery through technical means, including processes and results. However, in the traditional mode, there is a lack of process data, so it is difficult to carry out effective process measurement and analysis.

The benefit and effect are different.

DevOps supports continuous optimization for the delivery process. Through metric data analysis, we can quickly locate the problems and efforts of the delivery process, and support continuous scientific optimization, which is difficult to achieve in the traditional model.

DevOps landing value

The definition and evolution of DevOps and the comparison with the traditional model can see the systematic and integrated design idea of DevOps system itself, which includes not only the systematic consideration of the whole life cycle of software, but also the multi-demands of IT management. The landing of DevOps involves the construction of DevOps platform, process system construction, personnel empowerment, standards and norms, which can bring the following aspects of improvement.

Improved business responsiveness

The landing of DevOps can improve the quality of delivery while improving the efficiency, the improvement of the degree of automation can improve the response to business requirements iteration, and the throughput of R & D can be geometrically improved.

Improvement of R & D delivery specification

The R & D teams of different projects or products carry out R & D delivery activities on the same platform, and restrict the delivery activities of different R & D teams through the pre-established processes and rules of the platform, so as to achieve the unity of R & D delivery norms and standards. To achieve enterprise-level optimization, improvement and improvement.

Improve the efficiency of R & D delivery

Through the automation of environment preparation, compilation and construction, code quality inspection, system testing and software deployment in the process of R & D delivery, we can reduce the time of manual operation or wait for manual operation, comprehensively improve the automation level of R & D delivery process, and improve the efficiency of R & D delivery.

Improvement of R & D delivery quality

The quality of R & D delivery extends from the existing deployment result quality assurance to source code quality assurance and test coverage assurance, so as to realize the quality inspection and improvement of the whole process from source code to deployment, and improve the quality of R & D delivery through the whole link.

Improvement of R & D delivery control

Carry out comprehensive data and measurement according to the requirements, development, testing, release and deployment of R & D delivery, establish quality access control for key indicators, so as to achieve automatic technical control, combined with existing administrative control, by a single result control, achieve R & D delivery "process + result" management and control, and improve the intensity of management and control.

Continuous optimization of R & D delivery

Based on the DevOps platform, through pipeline process data collection and further measurement analysis, the continuous optimization of R & D delivery process is realized, including not only the optimization of DevOps platform, but also the optimization of R & D delivery process, standards and specifications.

Summary

The evolution of DevOps is simply from a technology to the construction and practice of culture; from irrelevance to dependence; from part of the business to the whole. The landing value of DevOps is realized in the integration of research and transportation and enhance the response to business needs. The trend has become. I only hope that everyone can participate in the event.

Author: Huang Quanzhong

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.