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

The little ant said:

Ant Financial Services Group has accumulated rich experience and best engineering practices in the continuous delivery of financial-grade Internet products. At the 2018 ATEC Technology Exploration Conference, Lu Zhongbang (Fengqi), Ant Financial Services Group solution architect, analyzed the core challenges of continuous delivery of financial-grade Internet products from the industry background, and shared the best engineering practices for ants to deal with these challenges from the two dimensions of "delivering value faster and earlier" and "keeping the technical risk bottom line to ensure delivery quality". Finally, the practical experience of ant R & D efficiency platform supporting continuous delivery is also introduced. Let's study with the little ants.

I. Industry background and main challenges

Under the background of digital transformation, enterprises need to build a variety of core competencies, which objectively require enterprises to upgrade or adopt a new generation of technological architecture. One of the most important aspects is the continuous delivery under the cloud-based infrastructure and distributed architecture. When it comes to continuous delivery, it is easy to think of some specific challenges, such as how to shorten the R & D and production time of new business products and quickly respond to customer needs; how to deal with the complex business scenarios and high concurrency challenges brought about by the distributed micro-service architecture; how to promote automation through technical means to reduce human investment in the R & D process, and so on.

In addition, we also need to seriously examine the characteristics of the industry. The two core keywords of financial Internet products, the first is "finance". The most important financial attribute is to ensure capital, security and high availability, which can be summed up in one word-"stable"; another keyword "Internet" is characterized by the rapid delivery of value and supporting the rapid innovation of the business. let's boil it down to another word-"fast". Not only to be fast but also stable, this is the basic characteristics of the financial Internet industry, seemingly contradictory two aspects, indispensable.

When it comes to "steady" and "fast", how is Ant Financial Services Group doing? Share several actual data from the last fiscal year: online service availability rate-100%; application release more than 150 times a day; iterative average R & D cycle 5.8 days; test automation rate more than 80%; operation and maintenance automation rate more than 98%.

Based on the background of digital transformation and the basic characteristics of the industry, we believe that the core challenge of financial Internet products in the field of continuous delivery is: how to strike a balance between speed and stability? It can not only deliver value quickly and quickly, but also innovate steadily, keep the bottom line of technical risk, and continuously meet the requirements of regulatory compliance.

Agile delivery-- how to deliver value faster and earlier

This chapter is divided into four parts, the first is Lean R & D process customization and diversification branch and release strategy, mainly to solve our system or process how to adapt to different business scenarios, the correct path and posture is the basic premise of R & D delivery efficiency.

The second is functional service, efficient joint adjustment and problem diagnosis, these two parts mainly explain how to liberate human flesh and improve efficiency by means of technology or automation.

1. Lean R & D process customization

When it comes to process customization, many people will ask: on what basis do we customize the R & D process? We have a more effective method in ants, that is, grading according to application. The application rating mainly considers three factors: the amount of dependent service transfer, the amount of daily transaction funds, and the daily PV and UV. According to these three aspects, we define twelve different application levels from A1-C4, and then set the baseline research and development rules for each level of application. On top of the baseline rules, we also support each business to customize additional risk control measures.

To give two examples, one is the on-demand configuration process. Ant Financial Services Group's business is very complex and diverse, and some of the core business systems require high stability, and the corresponding technical risk prevention and control measures, testing and verification links will be relatively perfect; on the contrary, some new businesses or internal service systems will tend to be online faster and earlier, and the process will be lighter and more agile. The other is the choreographable and expandable pipeline. The component center of the performance platform defines many quality testing components, including third-party or business-built components, and arranges personalized pipeline templates for different businesses through the orchestration ability of the platform. Some applications are forced to do code review, and some applications need to go through CI automated testing or some special testing before they can be pushed downwards. Similar scenarios can be implemented through pipeline choreography.

two。 Diversify branching and publishing strategies

There are four main ways for ants to play with branching patterns and publishing strategies.

First of all, the daily release, we compare it to a regular train, suitable for the station's core business systems, applications with strong correlation between the scene.

The second is independent release, which is compared to a car, which can start whenever you want, which is suitable for scenarios with independent business domains and certain coupling and correlation between applications.

The third is single application release, which we compare to a motorcycle, which is suitable for scenarios where business is more independent and is completely decoupled from other businesses at the architecture level. The first three modes usually adopt the mode of branch development trunk release.

The last emergency release, which we compare to an ambulance, is suitable for the resolution of emergency business needs or online failures, usually in the mode of branch development branch release.

Through these four modes, all the business scenarios of ants are basically covered, and each business can find a match according to its own needs.

3. Functional service

After introducing the posture and path of agile delivery, let's take a look at automated efficiency.

Usually in an R & D iteration, many functional departments are involved. the traditional practice is for each functional team to conduct human flesh risk control based on experience review. For example, when developers have completed coding and self-testing, there will be functional teams such as security and risk control to check and review based on experience, and the "department wall" will seriously affect the efficiency of collaboration and delivery. In ants, the cooperation ways of each functional team are completely different. Instead of directly participating in the project iteration and undertaking the repetitive mechanical activities of iterative verification and audit, they are transformed into capacity output and automated tool construction to achieve functional servicealization, thus empowering the business development and testing team. This model greatly improves the efficiency of R & D cooperation.

4. Efficient joint modulation and problem diagnosis

The business scenario of financial Internet products is very complex, and building a project environment is a very time-consuming and labor-consuming thing. For example, a transaction link involves 20 applications, the general practice is to deploy each application in the iterative process of research and development, and finally form a joint tuning environment. The approach of ants will be different. First, we build a shared STABLE environment. In the R & D iteration, we only need to deploy applications with changes and modifications, and then associate all 20 applications together through sofarouter grouping ability to form a joint tuning environment. This not only greatly improves efficiency, but also maximizes the use of test resources. In addition, when the code is released and launched, the platform automatically updates the STABLE environment to ensure that it is up-to-date.

If a problem is found in the process of joint adjustment, how to locate and diagnose the problem in such a complex link is also very important. The developer can query the link diagram and sequence diagram through TraceID or transaction number to understand the call interaction information between applications intuitively and comprehensively. Combined with the business log, it is very easy to find the wrong application and locate the source of the problem.

Third, steady innovation-- keep the bottom line of risk and ensure the quality of delivery.

This chapter is also divided into four parts, in which technical risk assessment, quality built-in and test verification are carried out according to the logic of the development before, during and after the development. finally, we share with you how we keep the security bottom line and ensure information security.

1. Risk assessment of data enabling technology

There is no doubt that the prior technical risk assessment of development is very important. Ant technology risk assessment is mainly based on two inputs: the first is demand input, the second is governance analysis related data input and empowerment, the latter is more important to us. Developer students can easily obtain data such as application dependence, service invocation, message inspection, component control, code retrieval, etc., and comprehensively and accurately evaluate the technical risks brought by the change. Based on these data and analysis, it is easy to determine the risk response strategy and achieve effective closed-loop feedback.

two。 Built-in quality real-time closed loop

In the process of development-built-in quality and real-time closed-loop feedback to help developers get things right the first time. Within ants, we encourage gitflow-based best practices to submit code to project branches or trunks through MergeRequest rather than Push, giving code access control and CI testing a chance. In fact, all the nodes and components of the Pipeline pipeline are choreographable and extensible. After the code is submitted, after each component is executed, the platform will feedback the results in real time, and automatically update the iterative quality data to assist the development and testing students to control the quality risk.

3. Full-environment and business hierarchical verification

After the development, that is, the test verification part, we share two points: the first is full-environment verification, from the development of "integrated" pre-release "grayscale step by step close to and simulate the production environment, to ensure that there are no problems with production release. The second is business hierarchical verification, in each environment, there is a corresponding testing means. For example, stress testing is being done by many companies, but it is mostly based on the offline environment, while ants will do stress testing directly in the production environment to really achieve the high availability of the system.

4. Information security guarantee

In the last part of safe delivery, Ant has a complete system for ensuring information security: in the requirements design review stage, the architect will assess the business risk; in the development phase, first, the security of the SOFA framework itself is guaranteed, and secondly, we have a secure automatic scan every time the code is submitted, in addition to a special security test. Finally, after the system is online, we have a special monitoring and emergency mechanism for security.

Fourth, AntLinkE support of R & D efficiency platform

Earlier we introduced not only how to deliver agile, but also how to innovate safely, and then how the platform level of sharing tools supports the entire continuous delivery process.

1. Brief introduction of platform

What the ant R & D performance platform does can be summed up in three aspects:

DevOps- one-stop development integration continuous delivery

DevMind- Real-time Multi-dimensional data Analysis enabling R & D process

DevServices- provides efficient technical support and consulting services to developers.

two。 Big picture of product

The following is a big picture of our product, the top is the business supported by the platform and the value of delivery, and the bottom is the main capabilities of DevOps, DevMind, and DevServices. Today we focus on the middle product layer.

The first is continuous delivery, from the beginning of the creation iteration to the end of the release, to provide support for the entire R & D life cycle. The following are supporting sub-products: research and development collaboration management projects, iterations and requirements; code services provide code hosting, code search, CR and other capabilities; IDE is a unique product of ants, which can help developers do code scanning at the first time, and integrate and integrate with the Web side of the performance platform, so that developers do not have to switch workbench frequently to carry out development work. The test service supports test management and automated testing; problem diagnosis relies on distributed links and business logs to quickly locate and resolve problems. The bottom is R & D insight, in the R & D process, both the IDE side and the Web side of the R & D efficiency platform will precipitate a large amount of data, which are very valuable assets. We continue to optimize and upgrade the entire R & D system and organization through collecting, statistics and analysis of these data.

3. One-stop continuous delivery solution

One-stop continuous delivery of the solution is shown in the figure. The middle part is the R & D performance platform, which provides a continuous delivery engine from requirements to release, connecting all related capabilities and tools together; at the bottom is the application PAAS platform, with which the performance platform interacts with it through openAPI, opening up functions such as environmental management and environmental deployment. On the right is distributed middleware. Through R & D containers, the R & D performance platform uniformly manages the configuration of multi-environment middleware, which not only realizes the isolation between environments, but also realizes the automatic conversion and synchronization between environments; in addition, the R & D efficiency platform is connected with the technology risk prevention and control platform, and the measures of technology risk prevention and control are implemented in the process of research and development. In addition, the ant R & D performance platform inherently has the ability of open integration, which can connect the enterprise's own tool platform through components to maximize the value of existing assets.

5. Conclusion

Whether internal or external, Ant R & D efficiency platform has been adhering to and will continue to pursue such an original intention-to enhance the happiness of developers, improve the efficiency of enterprise innovation, let's redefine R & D together!

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.