Shulou(Shulou.com)06/01 Report--

I believe that the operation and maintenance of children's shoes will have such an experience, "participated in 7-24 hours on duty, resistant to the machine, hit the code, and occasionally have to play with the data." In some small companies, operation and maintenance children's shoes dare not be said to be versatile, but at least they can fill in wherever there is a hole. However, the classification of operation and maintenance work in large companies is relatively fine. for example, operation and maintenance focused on highly repetitive work, business operation and maintenance focused on basic operation and maintenance work, planning and operation and maintenance focused on operation and maintenance work project management and requirements planning and design, development and operation and maintenance focused on the development of operation and maintenance tools, and so on.

The operation and maintenance work in large companies is not only faced with the complexity of the division of labor, but also subjected to the impact of peak flow. So, how is the operation and maintenance work of large companies managed? This article will take BAT as an example to uncover the tip of the iceberg in the thinking of operation and maintenance of large Internet companies.

Tencent: service, standardization, automation and production

When there are only a few people in the operation and maintenance team, the working status of everyone may be "individual sweeping the snow in front of the door" and go to support wherever the fire is needed, but if the team is still in this working state, when the number of the team increases to hundreds or thousands, there could be a big problem. Therefore, Tencent's operation and maintenance service has a "four modernizations", that is, service, standardization, automation and production.

How to understand the "four modernizations"? Service refers to serving the business and tends to solve the pain points of products; standardization is the basis of automation, because the differences between business architectures are different, so the granularity and dimensions of standardization also need to be adjusted; automation, we must open up the isolated islands of information in the enterprise, only the realization of the automation of the whole process is the real automation. Production, operation and maintenance platform should also pay attention to user interaction, to the concept of products to build.

Baidu: the book is the same as the text, the car is the same track, and the line is the same.

Judging from the information published on the Internet, Baidu's operation and maintenance has gone through four stages, namely, script tools, basic operation and maintenance platform, open operation and maintenance platform and intelligent operation and maintenance platform. the core goal of operation and maintenance has also transitioned from the previous improvement effect to how to liberate the efficiency of operation and maintenance.

At the SACC 2017 conference, a lecturer from Baidu once revealed that "the three ideas that Baidu guides operation and maintenance are the same book, the same track and the same line." How to understand it? The book and the article refer to the "language" of consistent operation and maintenance, such as the definition of operation and maintenance application, service, computer room and cluster; vehicle on the same track refers to the "method" of consistent operation and maintenance, such as capacity expansion and traffic switching execution; line homotopy refers to the consistent operation and maintenance mode, such as fault diagnosis strategy, flexible scaling policy, traffic scheduling strategy.

The image above shows Baidu AIOps technology stack, which is reported to be used in Baidu's internal, public and private cloud customers.

Ali: universal "large and medium stage and small receptionist"

It seems that the "fight monster upgrade" route of the operation and maintenance of each company is very similar, and Ali has also experienced the process of scripting, instrumentation, DevOps, automation and then intelligence. At the DTCC 2018 conference, Xu Xiaofei, a technical expert from Ali, made an analogy between AIOps unmanned movement and self-driving, both of which can be divided into six stages, and each corresponding stage has something in common. The six stages of AIOps unmanned operation and maintenance are human flesh operation and maintenance, script / tool operation and maintenance, platform operation and maintenance DevOps, data operation and maintenance DataOps, highly intelligent operation and maintenance AIOps and fully intelligent operation and maintenance AIOps.

People who know Ali should have heard of the term "large and medium-sized platform and small receptionist". This is the technology strategy pursued by all technology-related departments up and down Ali, and operation and maintenance is no exception.

Ali has a big data operation and maintenance solution called Tesla, which runs through Alibaba's entire big data system and is responsible for the operation and maintenance support of all products from infrastructure to infrastructure platform to storage computing platform. The vertical system of operation and maintenance is also a hierarchical structure of SPI, and the IaaS layer of operation and maintenance is responsible for providing basic operation capabilities for operation and maintenance management; the operation and maintenance PaaS layer provides functional services and data source and control services for various operation and maintenance fields; and the operation and maintenance SaaS layer provides scene-oriented automatic intelligent operation and maintenance applications.

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.