introduction
In the past 10 years, cloud computing has evolved from a single IT service to a new generation of software architecture paradigm, which in turn empowers the innovation of enterprise management and production models. Cloud computing has also experienced a development stage from "resources to the cloud" to "deep use of the cloud".
In the cloud-native era, the application lifecycle management method has become a focus of attention of enterprises. At the LeaTech Global CTO Leadership Summit co-hosted by Tencent Cloud TVP and 51CTO, Huang Junhong, vice president of Tencent Cloud, delivered a keynote speech on "The Way of Full Application Lifecycle Management in the Cloud-native Era", from the background of the new era of digital transformation Cut in and discuss how to manage cloud-native applications to fully release the dividends of cloud-native technologies.
The content of this sharing mainly includes the following three aspects:
- In the new era of digital transformation, a new paradigm of cloud native construction;
- Layout cloud-native, release room for growth in R&D efficiency;
- Build an ecosystem together to help the digital transformation of the entire industry.
1. In the new era of digital transformation, a new paradigm of cloud native construction
Affected by the epidemic and the international environment, the growth of my country's cloud service market has slowed down. According to statistics in 2021, the scale of China's cloud computing market is about 300 billion. Although the growth rate reaches 48.4%, compared with 2020 It has been reduced by 8.3%. After experiencing a blowout period in 2020, the overall growth rate of the public cloud service market has slowed down, and it has entered a bottleneck period from the original high-speed growth period. The in-depth advancement of the public cloud market has ushered in new opportunities.
At present, my country has formed the basic digital economy of 11 large urban agglomerations headed by the Pearl River Delta. Starting from 2021, three urban agglomerations with particularly prominent growth have also appeared in the digital economic belt in the middle reaches of the Yangtze River. The average growth rate in the past two years has reached 124%, which can be said to have become an important source of growth for digital transformation. Across the country, the digital transformation of enterprises has also brought a steady stream of new vitality to the entire cloud market.
In the process of promoting the digital transformation of enterprises, it has become an inevitable trend to build a digital infrastructure based on cloud computing and realize the modernization of IT architecture. As the technical core of next-generation cloud computing, cloud native has become the best practice path to give full play to cloud efficiency with its unique advantages such as independence, high efficiency, and elastic expansion. With the gradual maturity of cloud-native technology and capabilities and the promotion of the new economy, enterprise IT elements are constantly changing, and cloud usage has gradually entered a new stage of development.
From the above data, it can be seen that the migration of IT infrastructure to the cloud in various industries in my country has become a mainstream trend, and the rate of enterprises in various industries going to the cloud continues to increase. In traditional industries such as industry, the current cloud deployment rate has reached 45%. In addition, the international authoritative organization Forrester also predicts that 50% of enterprise development teams will switch to an integrated DevOps tool chain in 2022. This year, we can also see the vigorous development of one-stop R&D management platforms and services, which indirectly confirms Forrester's previous forecast.
Tencent is a firm supporter of cloud-native technology and one of the earliest manufacturers to deploy cloud-native business. As early as 2015, when Docker and K8s were first launched, Tencent's social business and some game businesses began to use these two technologies. In 2017, based on its internal experience, Tencent launched container service TKE and cloud function service SCF for public cloud customers, and gradually launched other public cloud native services. In 2019, Tencent decided to fully move its massive internal self-developed business, including social networking, games, and content, to the cloud, and complete the cloud-native transformation. In 2022, Tencent's self-developed business will initially complete the transformation of cloud-native technology. The scale of self-developed business on the cloud has exceeded 50 million cores, and the cumulative cost savings have exceeded 3 billion.
On the other hand, the self-developed business on the cloud also allows Tencent to establish a future-oriented technical architecture and R&D system. Comprehensive cloud native not only improves our resource utilization, but also enables Tencent to have better elastic expansion capabilities, which can provide strong support for the rapid development of internal self-developed business. Using the rich products and capabilities on the cloud, the agility of business development has also been further improved. For example, the star products Tencent conference and video account in the past two years have all grown up on the cloud. Through Tencent Cloud's elastic scaling capabilities, Tencent Conference has supported tens of millions of PCUs, and the video account has also quickly launched many innovative functions with the help of cloud live broadcast PaaS capabilities. It has been reduced by 30% before, and the improvement effect is very obvious.
2. Layout cloud native to release room for growth in research efficiency
Tencent Cloud's cloud-native layout can be said to be full of flowers, from computing cloud-native, architecture cloud-native, security cloud-native, to upper-level development of cloud-native and data cloud-native, etc. Among them, cloud-native application management is Tencent Cloud's development in the field of cloud-native important part of the layout.
(1) Challenges in the cloud-native era
As a technical manager, the challenge encountered in application management is often the problem that the dividends of cloud-native technology cannot be fully released when implementing cloud-native technology. The reason behind this is often that each team uses cloud-native technology independently according to their own understanding and knowledge, lacking unified standards and specifications, and after a period of development, many technology stacks that are separated from each other will be formed. R&D, delivery, and operation and maintenance are not standardized enough, and enterprises also lack the support of tool systems to coordinate with the implementation of standards. As a result, the published standards and specifications can only be suspended in the air and difficult to implement. Ultimately, the effect of going to the cloud is not good.
For R&D personnel, this challenge often comes from the complexity and mental burden of cloud native technology.
The left side of the picture above is CNCF's cloud-native landscape. From 2017 to 2022, the number of applications has risen sharply. The CNCF 2022 survey report also shows that nearly 55% of the teams believe that the biggest difficulty in going to the cloud is lack of training. This means that in order to make good use of cloud native, R&D needs to continuously learn cloud native technology. The complexity of cloud native technology is a test of the learning ability of R&D personnel. When this complexity rises to a certain level, R&D will be too busy dealing with various new terms and new technologies to take care of business development, unable to efficiently focus on the creation of business value, and it will also lead to an increase in organizational training and recruitment costs.
(2) Tencent Cloud's native application and management approach
In order to solve the above problems, Tencent's cloud-native application management approach is centered on the application. On the premise of ensuring the security of resources, it provides R&D personnel with automation tools, so that they can complete the full application of the application in a low-threshold, efficient and reliable manner. Lifecycle management means that R&D personnel can independently and declaratively complete infrastructure delivery, simple application modeling, and efficiently and reliably complete application delivery and application O&M.
As a cloud operating system, K8s must be complex, but we can solve this complexity problem by separating perspectives. The organization's architecture team, as cloud-native experts, builds organization-level cloud-native technical standards through templates around the business, unifies the delivery of cloud-native capabilities, and shields most of the complexity of cloud-native modeling. Through cloud-native standardized governance and shielding the complexity of cloud-native, organizations can fully release the dividends of cloud-native technology, and also allow R&D personnel to return to the creation of business value and improve organizational efficiency.
In the cloud-native era, infrastructure and applications are maintained declaratively. A declarative approach to delivery is a more reliable approach to process-oriented delivery. The process-oriented tools of the old days can no longer take full advantage of the reliability of the cloud. There are many examples of production accidents caused by improper operation, wrong and missed delivery, configuration conflicts, configuration drift, etc. in production and research. Through the practice of advanced concepts such as GitOps and Application as Code, we have established a declarative delivery capability to solve the efficiency and reliability issues of multi-environment and multi-cluster delivery for organizations, thereby improving cloud-native delivery efficiency and delivery quality.
The cloud-native application of enterprise applications makes traditional monitoring problems such as isolated tool islands, split perspectives, and data faults. The establishment of an application-centric hybrid cloud unified observation plane has become the primary goal. To put it simply, it takes business as the starting point to establish an end-to-end full-stack observability. There are two dimensions here. One dimension is the access flow dimension, such as from the terminal, access layer, data layer, etc. This is the data request flow; another dimension is the application layered dimension, such as from the application layer to the operating system layer to the protocol stack to the network, which is the vertical dimension. Through this multi-dimensional data collection, unified modeling and analysis, we can achieve full-link performance monitoring, intelligent alarms, root cause analysis, anomaly detection, fault self-healing and other goals, and truly realize the transformation from monitoring to observability comprehensive upgrade.
In terms of infrastructure delivery, in 2022, HCL, which orchestrates infrastructure through IaC, will become GitHub's fastest-growing language. The reason behind it is that the artificial "bit by bit" method cannot meet the needs of large-scale cloud infrastructure management. IaC arranges the infrastructure through code, and cooperates with GitOps to realize the automatic management of infrastructure, which greatly improves the management of cloud infrastructure. efficiency.
While introducing IaC capabilities, Tencent Cloud CODING has strengthened the link between enterprise applications and cloud products, enabling enterprises to establish associations between business, applications, and infrastructure from a business perspective, and establish a cost management foundation from a business perspective for enterprises. While improving the efficiency of infrastructure management, it improves business IT cost insight capabilities and further empowers businesses from a technical perspective.
At present, Tencent Cloud CODING DevOps has been online and running stably for 8 years, accumulating more than 3 million developer users, 50,000 enterprise teams, and serving customers in different industries such as Internet, finance, government and enterprises. The Orbit cloud-native application management platform self-developed by Tencent Cloud has also accumulated a lot of practice in the financial, medical, retail and other industries.
3. Build an ecosystem together to help the digital transformation of the whole industry
Relying on Tencent's more than 20 years of technical training and service experience in serving billions of massive users, Tencent Industrial Internet uses the cloud as the base, combining leading technological capabilities such as audio and video, big data, and corporate WeChat, Tencent Conference, and Tencent Documentation and other enterprise service applications, and create professional solutions for different industry characteristics. At the same time, fully open internal ecology such as content numbers and small programs, and through Tencent Yunqi, Tencent Qianfan and other platform plans, cooperate with external ecological partners to provide more than 300 industry solutions for more than 20 industries, stimulate new efficiency of the real industry, and start the digital economy The new engine helps all walks of life to achieve digital upgrades.
In the process of helping the industry, Tencent Cloud has always maintained enthusiasm and openness for technology. It has established Tencent Cloud TVP as an in-depth communication platform with industry technical experts, and regularly holds activities such as cutting-edge technology discussions, in-depth product experience, and offline visits and exchanges. At the same time, Tencent Cloud has also launched the Tencent SME Digitalization Level Evaluation System to help enterprises understand their own digital development level, so as to better grasp the direction of transformation.
Finally, talk about some prospects for the future. One of the top ten technology trends released by Gartner in 2023 is platform engineering. Gartner believes that in the context of cloud native, there should be a service-oriented DevOps concept, and it is necessary to provide DevOps support for service teams through platform-level capabilities.
The R&D support team should provide the development team with an engineering platform to help the development team shield the complex underlying infrastructure. The corresponding product capabilities are roughly divided into automation platform, cloud-native construction, DevSecOps, and cloud-native application management.
It is believed that in the future cloud-native era, DevOps can bring more productivity improvements to enterprises and further promote the transformation of industrial digitalization.