Break through with operation and maintenance tools! See how the operation and maintenance department of China Merchants Bank gets along well with development and business

[51CTO.com original article] When traditional banks begin to transform into Internet finance, banks not only need to reshape their own image, but also use various means and tools of Internet+ to become more financial service providers. As the saying goes, "a good iron must be made strong", the IT architecture and system of traditional banks also need to "transform" urgently to carry more business value and help business departments build a more complete digital ecosystem.

Zhang Jianlin, head of the data center application management team of China Merchants Bank, recently talked with reporters about the evolution of non-functional practices of traditional financial IT. He shared with reporters the evolution of traditional financial needs under the trend of Internet finance, the linkage between the operation and maintenance department and other development departments and business departments, and how to make operation and maintenance work keep pace with the times and play more value in an era of emerging technologies.

[[201216]]

Four major challenges of traditional finance

Zhang Jianlin divides the challenges that traditional finance poses to operations and maintenance into four categories:

First, there is the pressure from the rapid growth of online business transaction volume. The original organizational structure can no longer meet the demand for such high transaction volume.

Secondly, the operation and maintenance department and the development department need to be seamlessly connected in their work. He told reporters that due to policy requirements, many operation and maintenance tasks cannot involve developers, and there is a clear division of labor between development and operation and maintenance. However, for Internet companies that are developing at a high speed, it is particularly important to ensure that these two departments can achieve seamless supervision;

Another problem is that the growth rate of operation and maintenance personnel can never keep up with the growth of development and business demands;

*** is the requirement of bank regulatory authorities and bank customers for high availability of application systems. This fully reflects the particularity of the financial industry. Unlike the high availability of general enterprises, the business systems of banks need to meet the high standards of the China Banking Regulatory Commission and the China Securities Regulatory Commission. The security of data and transaction security and availability must be considered first, followed by how to cope with high concurrency requirements. These have brought considerable challenges to banks.

Operation and maintenance work also has pressure

At the 2017 China Application Performance Management Conference hosted by Tingyun, Zhang Jianlin gave a speech titled "The Evolution of Traditional Financial IT Non-Functional Practices". When the reporter asked what is the "non-functional IT" of traditional finance, Zhang Jianlin explained that the needs of the business department are called functional requirements, and the purpose is to achieve a certain business need. Non-functional requirements are requirements raised from the perspective of operation and maintenance, mainly in order to achieve the demands of the project in addition to completing the development of business functions. He gave an example. For example, if the business department proposes a flash sale activity, then the first task of the development department is to design a program that meets this flash sale activity, but they will not consider how much storage, network, and computing resources the IT system needs to be equipped with behind this activity, whether stress testing is required, whether performance tuning is required, what is the environment deployment of this flash sale activity, etc., and these factors are exactly what the operation and maintenance department needs to consider. Then the operation and maintenance department will raise IT non-functional requirements in the code development stage, requiring the development department how to configure the database and how to achieve load balancing and matching business. Ensure that development and operation and maintenance jointly achieve business needs.

Zhang Jianlin pointed out that traditionally, business departments did not communicate much with operation and maintenance departments, but now the operation and maintenance departments are more actively involved in the work of the business departments. They will reach a consensus with the business departments in advance and fully understand the business scenarios, so as to have an estimate of the business volume and determine the performance bottlenecks, and make environmental deployment in advance, rather than waiting until the business volume increases suddenly, and then expand bandwidth and buy equipment to avoid a passive situation.

Breaking the deadlock through self-reliance

Zhang Jianlin said that there are two major schools of thought in operation and maintenance. One is DevOps, which advocates that applications, operation and maintenance are integrated and seamless with development. The other is the SRE concept (Site Reliability Engineer) proposed by Google, which allows development and operation and maintenance to work together. Operation and maintenance can intervene in project development in advance and prepare some ideas or tools, scripts, and interfaces for operation and maintenance in advance for development. "I agree more with the concept of Google SRE, which is more feasible in reality." Zhang Jianlin said frankly.

Another powerful tool to solve the operation and maintenance deadlock is automation. Zhang Jianlin said that under the current X86 platform, it is impossible for one person to operate and maintain one or two thousand servers, so automation is necessary to make a smooth and seamless connection between development code and operation and maintenance automation.

The reporter learned that at China Merchants Bank, all automation tools are independently developed by the operation and maintenance team, because the independently developed automation tools are more in line with the unique applications, unique environment, and unique staffing of China Merchants Bank. At present, the operation and maintenance and data center departments of China Merchants Bank can fully carry out operation and maintenance work with these tools, support the applications delivered by the development department, and ensure the high availability of applications. "You are the one who knows your needs best, and the tools you make may be the most suitable." Zhang Jianlin said that some manufacturers were willing to provide customized automation products before, but it was difficult for outsiders to deeply understand the environment and business of the bank, and the response speed and intensity could not keep up. The continuity of product iteration was not very good, and it was easy to cause waste. These can be effectively compensated by independently developing automated operation and maintenance tools.

The next focus of operation and maintenance

In the interview with ***, Zhang Jianlin also talked about his views on the future focus of operation and maintenance work. He believes that there are three changes:

First, traditional financial institutions are also gradually catching up with Internet companies, gradually strengthening the collection of online customer behavior data, doing big data analysis and mining, and mining data value from the business level. The responsibility of the operation and maintenance department is to provide these business departments with different data sources and big data analysis platforms.

Another obvious trend is that the focus of traditional finance is gradually shifting from offline to online, but the previous application architecture and capacity design certainly cannot meet the current needs and development of online transactions, especially in terms of high concurrency, high transactions and high growth. Therefore, it is now necessary to gradually rectify and migrate the previous old architecture, mainly by decoupling the application of such business and moving it down to the X86 platform from the host.

Zhang Jianlin said that the data center will accelerate the construction and operation and maintenance of the cloud platform in the future. At present, some non-core businesses have been put into production and operation on the private cloud. When the time is ripe, more businesses will be migrated to the cloud platform for deployment.

APMCon2017

APMCon is a global high-level APM technology event jointly organized by Tingyun, GeekBang Technology and InfoQ, focusing on the most critical performance monitoring and management technologies for mobile, Web and Server terminals. The entire conference includes topics such as intelligent operation and maintenance, database performance optimization, mobile performance optimization, front-end performance optimization, real-time analysis and visualization of business data, microservices, containers and CDN acceleration.

APMCon content is based on practice and is oriented towards the community. Speakers from home and abroad share APM technology innovations, trends and best practices based on hot topics with technical team leaders with more than 5 years of experience, mid- to senior-level developers and operation and maintenance personnel, and engineering directors.

[51CTO original article, please indicate the original author and source as 51CTO.com when reprinting on partner sites]