The following is the China Cloud native survey in 2020 From CNCF recommended by recordtrend.com. And this article belongs to the classification: cloud computing.
CNCF conducts community research regularly to better understand the application of open source technology and cloud native technology. CNCF conducts regular community surveys to better understand the adoption of open source and cloud native technologies. For the fourth time, we conducted a “cloud native survey in China” in Chinese to understand the speed of cloud native adoption in China.
The survey was completed in December 2020 and January 2021, and 91% of the 439 respondents came from Asia. The survey was conducted in December 2020 and January 2021 respectively, and 91% of the 439 respondents came from Asia.
Main excerpts from China Cloud native survey:
Containers continue to grow rapidly. 68% of institutions use containers in the production process, an increase of 39% over last year and an increase of 240% two years ago. Containers continue to grow dramatically. 68% of organizations use containers in production, an increase of 39% over last year and 240% over two years ago.
Kubernetes are ubiquitous, and the proportion of kubernetes used in production has increased from 72% last year to 82%. Kubernetes became ubiquitous. Kubernetes use in production has increased from 72% last year to 82%.
Public cloud is back to the top, with 46% of respondents using public cloud, higher than 36% last year, but lower than 51% in 2018. 46% of respondents use public cloud, up from 36% last year, but down from 51% in 2018.
The degree of automation continues to improve. 30% of the respondents used automatic release cycle, and 73% ran CI / CD pipeline. Automation is improving. 30% use automatic release cycle and 73% use CI / CD pipeline in production.
Observability tools are widely used. 95% of units use monitoring tools, 94% use logs, and 85% use distributed tracking. Observability tools are widely used. 95% of organizations use monitoring tools, 94% use logging, and 86% use distributed tracking.
The use of containers continued to increase significantly. Container usage has grown in all application categories this year, including testing, proof of concept (POC), development, and production. Currently, 94% of respondents use containers at least one stage. Among the respondents who used containers, 68% used them in production, an increase of 39% over the previous year. Since we began to survey Chinese communities in 2018, the proportion of containers used in the production stage has increased by 240%. The use of containers is increasing dramatically. This year, it has been added in all categories: testing, proof of concept (POC), development and production. Now 94% of respondents use containers at least one stage. Of those who use containers, 68% use them for production, up 39% from last year. Since we started our survey of Chinese communities in 2018, the number of containers used in production has increased by 240%.
The percentage of units using containers in test environments has increased by 150% last year, which means that the use of containers in development and production will continue to grow. The number of container users in the development phase has increased from 34% last year to 72%, and the number of container users in the POC phase has increased from 21% to 60%. Compared with organizations that use containers in test environments, it also grew by 150% last year, which means that container growth in development and production will continue. The number of organizations using containers in development has jumped from 34% last year to 72%, and POC has jumped from 21% to 60%.
The number of containers used by each unit is also increasing. Last year, 31% of respondents said their units used less than 50 containers. The situation has changed this year, with 31% saying they have used more than 5000 containers per unit. Overall, about half (48%) of the units use 1000 or more containers. Organizations also run more containers. Last year, 31% of respondents said their organization used less than 50 containers. This year, that has changed, with 31% saying their organization uses more than 5000. Overall, nearly half (48%) of organizations are running 1000 or more containers.
When users are looking for a higher degree of automation in the deployment process, CI / CD becomes the second biggest problem after complexity. In terms of containers, complexity is listed as the primary challenge of using and deploying containers for three consecutive years. Next is CI / CD, because organizations want to add more automation to their deployment.
82% of the units used kubernetes in production, up from 72% in the last survey. Kubernetes has become the de facto container coordinator in China, of which 82% organizations use it in production. This is an increase from 72% in the last survey.
The number of kubernetes used by each unit increased at the same time. The easiest choice for respondents is to choose 2-5 kubernetes for production growth. This year and last year, 35% of users chose 2-5 kubernetes for production utilization. The number of units using more than 50 containers also increased, from 13% in 2019 to 21%, an increase of 320% over 2018. Organizations also use more kubernetes clusters. Among the two to five kubernetes production clusters, the respondents were the most popular choice, with a selection rate of 35% this year and last year. The number of containers using more than 50 containers has increased, from 13% in 2019 to 21%, an increase of 320% over 2018.
Separating kubernetes applications
Of the units using kubernetes, 84% use namespace separated applications, up from 71% in 2019. 58% uses split allocation, up from 42% in 2019. In organizations using kubernetes, 84% of respondents used namespace to separate applications, up from 71% in 2019. 58% of respondents used separate clusters, up from 42% in 2019. 26% of respondents only used tags, the same as last year.
Among organizations with multiple teams, 80% use a namespace to separate kubernetes applications, up from 68% last year. 60% of them used the allocated share, up from 46% last year; Another 22% use only tags, and in organizations with multiple teams, 80% use a namespace to separate kubernetes applications, up from 68% last year. 60% of them used a single cluster, higher than 46%; 22% only use labels, less than 25%.
The kubernetes package became the choice of 64% of users, up from 54% last year. Again, helm is the most popular kubernetes packaging method, with a packaging rate of 64%, up from 54% last year.
During local container development, 37% of respondents said they would like to choose docker kubernetes. Compared with last year’s ranking, this figure is down from 42% last year. Kubernetes of cloud vendor management became the choice of 28% of the respondents, up from 14% last year, and surpassed minikube, which ranked second last year. This year, 21% of the respondents chose minikube, ranking third. In the local container development process, 37% of the respondents said they targeted docker kubernetes. This is down from 42% last year. Kubernetes, managed by cloud providers, came in second, rising to 28% from 14% last year. It overtook last year’s second place minikube and is now third with 21%.
Workload auto scaling
Among the users who plan to scale kubernetes workload, 69% expect to scale stateless applications automatically, up from 54% last year; 35% of them plan to scale tasks / take measures, up from 26% last year; 34% plan to scale automatically. For those who plan to expand kubernetes workload, 69% plan to scale stateless applications automatically, up from 54% last year; 35% plan to expand task / queue processing, up from 26% last year, and 34% plan to automatically scale stateful applications, up from 23% before.
Of the users who don’t use kubernetes auto scaling, 31% have established their own solutions and 28% use third-party solutions. In the last survey, 28% of users didn’t want to auto scale any work, but the more they could. Among users who didn’t use kubernetes auto scale, 31% built their own solutions, while 28% used third-party solutions. In the last survey, 28% of people did not want to auto zoom any work, but as more and more organizations use kubernetes in production, the proportion dropped to 13%.
Cloud and code
All units submit codes more frequently. 43% of the respondents said that their units submitted codes once a day, an increase of 0.31% compared with 35% last year, and submitted codes several times a week, a decrease compared with 43% last year. In other words, nearly three-quarters of organizations now submit code at least several times a week. Organizations check more often in code. 43% of respondents indicated that their organization checked the code several times a day. This is up from 35% last year. 31% checked in code several times a week, down from 43% last year. This means that nearly three-quarters of organizations now check code at least a few times a week.
The release cycle of each unit was further extended. The number of units that need to be released every day or more substitutes has increased from 18% last year to 27%. Nearly two-thirds (62%) of users publish software every week, even more repeatedly. Organizations have more frequent release cycles. The number of people released every day or longer increased from 18% last year to 27%. Nearly two thirds (62%) release software weekly or more frequently.
Automatic and hybrid release cycles are becoming more popular. The number of users using manual release cycles has dropped from 31% last year to 17%. Automatic and hybrid release cycles are becoming more and more popular. The number of users using manual release cycles has dropped sharply from 31% last year to 17%.
Mixed release cycle, the proportion of this method increased slightly from 46% last year. The number of users using automatic release cycles has increased from 21% to 30%. More and more companies continue to use CI / CD tools, and it is expected that more users will use the automatic release cycle in the future. Less than half of organizations have hybrid release cycles, the most popular method, up slightly from 46% last year. The number of people using automated release cycles has increased from 21% to 30%, and we expect this growth to continue as more and more organizations continue to use CI / CD tools.
Public cloud ranked first, with 46% of respondents taking it as their preferred data center category. That’s up from 36% last year, but still less than 51% in 2018. Another 45% of users use private cloud and 42% use hybrid cloud, both of which are slightly larger than last year. This survey is the first time to understand the use of “multiple clouds”, and the results show that 22% of the respondents chose multiple clouds. In the future, as more and more units seek to use the best solution, hybrid cloud and multiple clouds are expected to achieve growth in use. The public cloud is on the back, but mixed and cloudy are ready to grow. Public cloud usage ranked first, with 46% of respondents calling it the top data center type. This is up from 36% last year, but down from 51% in 2018. Private (45%) and mixed (42%) also showed a slight increase compared with last year. The survey asked “cloudy” for the first time, and 22% of respondents chose it. As more and more organizations want to leverage best in class solutions, we hope to see growth in future mixed cloud and cloudy surveys.
Please note that allowing respondents to choose multiple answers may lead to overlap. We will address this issue in future investigations.
The number of users using more than 5000 machines has increased. However, the number of users using the most is 6-100 units. This year, 47% of the users in this range, compared with 44% last year. Only 15% of the units use more than 5000 machines, but the proportion is up from 10% last year. The number of machines that are currently in use exceeds 5000 tons is increasing. However, the most popular number was between 6 and 100 machines, with 47% of respondents pointing out that, compared with 44% last year. Only 15% of organizations have more than 5000 computers, but that’s up from 10% last year.
Cloud native tools
Building cloud native infrastructure can bring many benefits. Respondents ranked the benefits of cloud native tools: Although there are many benefits of building cloud native infrastructure, respondents ranked the biggest benefits of cloud native tools as follows:
Shorter deployment time (59%), shorter deployment time (59%)
Improved scalability (55%), scalability (55%)
Cloud portability (50%) cloud portability (50%)
Developers increase (49%) productivity (49%)
Increased availability (47%) availability (47%)
CNCF project CNCF project
The use of CNCF projects in Chinese communities is increasing.
Last year, the use of graduation CNCF projects increased by 64%, and the use of several projects in production showed a leap forward growth. In particular, some projects show impressive leaps in production and use
The number of users using Prometheus in production increased from 57% to 75%, an increase of 32%. The number of users using Prometheus in production increased from 57% to 75%.
Coredns usage increased from 35% to 64%, an increase of 83%. Coredns usage increased from 35% to 64%, an increase of 83%.
Etcd usage increased from 37% to 77%, an increase of 108%. Etcd usage jumped from 37% to 77%, an increase of 108%.
The first graduation project from China, harbour, increased from 27% to 47%, an increase of 75%. Currently, there are two graduate programs from China: harbor and tikv. Harbor is the first graduation project in China, with 75% increasing from 27% to 47%. Now there are two graduation projects in China, harbor and tikv.
We lack data that does not include fluent data.
Since last year, the use of incubation programs has increased by 94%. Accelerated growth projects include: the use of incubation projects has increased by 94% since last year. The fastest growing projects include:
CNI increased by 141%, and the number of respondents using the project in production has increased from 22% to 53%. CNI grew 141%, from 22% to 53% of respondents using it for production.
The number of grpc used in production has increased from 29% to 47%, with an increase of 62%. Grpc, whose production and usage increased from 29% to 47%, an increase of 62%.
23% of the respondents used opentracing and crio-o in production. Opentracing and cri-o were used by 23% of respondents.
Since contour, cortex, emissary express and operator framework were added after the survey, the data of the above items were omitted in this survey. The investigation has begun.
Since the last report, TOC has changed the process of accepting sandbox projects, so far we have seen a large number of projects join. At present, we have 50 sandbox projects! Part of the project in the production of the use has reached two numbers. In the last report, TOC changed the process of accepting sandbox projects, and we saw a massive influx. Now, we have 50 sandbox projects! Some even use double-digit numbers in production.
18% of the respondents have used virtual kubelet in production, making it the most used sandbox project. Virtual kubelet is the most commonly used sandbox project with 18% utilization rate in production.
Two service grid projects followed closely, network service grid won 15% users, and service grid interface won 13% users. Next, there are two service grid projects. Network service grid accounts for 15% and service grid interface accounts for 13%.
CI / CD
73% of respondents are running CI / CD pipeline in production, another 19% are evaluating, and another 15% plan to use automatic release cycle in the next 12 months. As mentioned above, more and more organizations are using automatic release cycle. 73% of respondents run CI / CD pipeline in production. Another 19% are in the process of evaluation, and another 5% plan to use them in the next 12 months.
Jenkins is the CI / CD tool with 52% users; Yes, another 35% use gitlab and 17% use Argo. Jenkins is the most popular CI / CD tool in the Chinese community. 52% of organizations use the software, 35% use gitlab, and 17% use Argo.
Monitoring, logging and tracking
With the increasing importance of cloud native observability, more and more attention has been paid to monitoring in the process of deployment. As the deployment becomes more and more complex, the visibility is more important than ever. As organizations realize the importance of cloud native observability, they increasingly run monitoring, logging, and tracking tools in their deployment.
95% of the units use monitoring tools, 94% use log, 86% use tracking. Most of the companies using the tools will choose to run locally on their own infrastructure instead of using managed services through remote services. 95% of organizations use monitoring tools, 94% use logging and 86% use tracking. Most of the people who use these tools run them locally in their infrastructure, rather than hosting them through remote services.
No server no server
At present, 31% of the units use no server in production, 41% are in evaluation, and 12% plan to use it in the next 12 months. There is still room for growth in serverless, but we expect that growth to continue. 31% of organizations use no server in production. 41% are evaluating – 12% plan to use it in the next 12 months.
Among the users without server, 45% use managed platform and 20% use installable software. Among the servers without server, 45% use managed platform and 20% use installable software.
The solution is favored by 35% of users, and another 19% of users use AWS lambda. In the organizations that use no server through the “hosting platform”, the most popular solutions are again 35% of alicloud functions and 19% of AWS lambda. Overall, with the emergence of new solutions, the response is more decentralized.
Among the organizations that use no server through installable software, 31% will use knative, up from 22% in the last survey. Knative surpassed kubeless, and kubeless usage dropped from 29% to 11%. Virtual kubelet, a sandbox project, ranks third with 9% ownership. Among the organizations that use no server through installable software, 31% use knative, compared with 22% in the last survey. Knative overtook kubeless, which fell from 29% to 11%. Virtual kubelet, the sandbox project, ranked third with 9%.
Service grid is very popular, 38% of the units use service grid in production, but its complexity still needs to be reduced. 42% of the respondents are evaluating the service grid, and 11% plan to use it in the next 12 months. Therefore, it is expected that 38% of the organizations that use the service grid in production still have complexity in solving these problems, although the service grid is very popular. 42% of respondents are evaluating the use of service grid, while 11% plan to use it in the next 12 months, so we expect this to be the growth area in the next few years.
In which service grids are the most popular tools
At present, most (56%) units use service agents in production. Another 26% are being evaluated and 7% plan to use them in the next 12 months. Most organizations (56%) now use service agents in production. Another 26% are evaluating and 7% plan to use one of them in the next 12 months. As service grid usage grows, we expect entry tools to grow as well, as they help control traffic in and out of the service grid.
Nginx (F5) has once again become a tool for translation, but its usage has decreased from 54% to 39%. Envoy, which ranked fourth last year, surpassed F5 and haproxy, with their share increasing from 15% to 29%. Nginx (F5) is the most popular tool, but it has dropped from 54% to 39%. Envoy, which ranked fourth last year, grew from 15% to 29%, surpassing F5 and haproxy.
Cloud native storage tools are also very popular, with nearly two-thirds (64%) of units using stateful applications in production. Another 21% are being evaluated and 5% plan to use them in the next 12 months. Cloud native storage tools are also popular, with nearly two-thirds (64%) of organizations running stateful applications in production. Another 21% are in the process of evaluation, and another 5% plan to do so in the next 12 months.
In Yunyuan, 31% used CEPH, 26% used CSI, 15% used gluster and 15% used Minio. For cloud native storage, 31% uses CEPH, 26% uses CSI, 15% uses gluster, and 15% uses Minio.
Cloud native communities in China
At present, CNCF has 60 members in China, accounting for more than 8% of the total number of members. After the United States and Germany, China is the third largest contributor base (contributors and contributors) of CNCF project. Among all members, pingcap and Huawei are the sixth and eighth largest contributors with nearly 85000 and more than 66000 contributions respectively. CNCF currently has more than 60 members in China, accounting for more than 8% of the total number of CNCF members. China also has the third largest donor base of CNCF projects (contributors and contributors) after the United States and Germany. Members pingcap and Huawei are the sixth and eighth largest contributors respectively, contributing more than 85000 and 66000 respectively.
CNCF has 11 CNCF projects from China: BFE (Baidu), chaos mesh (pingcap), chubaofs (Jingdong), CNI Genie (Huawei), dragonfly (Alibaba), harbor (VMware China), kubeedge (Huawei), openkruise (Alibaba), openyurt (Alibaba), tikv (pingcap) and volcano (Huawei). CNCF also hosts 11 CNCF projects born in China: BFE (Baidu), chaos mesh (pingcap), chubao fs (Jingdong). COM), CNI Genie (Huawei), dragonfly (Alibaba), harbor (VMware China), kubeedge (Huawei), openkruise (Alibaba), openyurt (Alibaba), tikv (pingcap) and Volvo (Huawei).
In China’s cloud native communities, 94% said they would recommend CNCF projects and technologies.
Learn more about cloud native
China’s cloud native community will understand cloud native technology through a variety of ways.
Podcast and webinar podcast and Webinar
Technology podcast is the most targeted response about how the community understands cloud native technology, and 54% of the respondents chose technology podcast.
Recently, at the request of the community, we have modified our online programs, and now we can provide live and on-demand services for webinars. Live and on demand webinars. We also have a weekly live community broadcast called cloud native live. Regularly arrange programs for Chinese audiences.
Case studies are also popular. 39% of the respondents chose business case analysis and articles, and 31% chose kubernetes case analysis. Case studies are another popular answer. The selection rate of business case studies and articles was 39%, and that of kubernetes case studies was 31%.
In 2020, together with China’s end-user members, China Minsheng Bank and Jingdong group, we released two new CNCF case studies. In 2020, we launched two new CNCF case studies for Chinese end users, including China Minsheng Bank and Jingdong com。
45% of the respondents chose kubecon + cloudnativecon to learn about cloud native technology. 45% of respondents chose kubecon + cloudnativecon event as a place to learn cloud native technology.
In 2020, we hosted the first online conference “cloud native + open source virtual summit China 2020” in China, with 5800 participants. In the second half of 2021, we will hold “kubecon + cloudnativecon + open source summit China”. In 2020, we held our first virtual event in China, “cloud native + open source virtual summit China 2020”, which attracted more than 5800 participants. We look forward to the kubecon + cloudnativecon + open source summit China event later in 2021.
Another 29% chose parties and local activities. In 2020, CNCF has set up a CNCF community group as a community activity site to hold parties, kubernetes community day and other activities. Group or activity. Another 29% of gatherings and local events. CNCF launched the CNCF community group in 2020 as a home for community activities, including meetups and kubernetes community day. Looking ahead, we want these events to have physical and virtual components. Find groups or activities near you.
More than 40% of the respondents have access to cloud native technology through kubernetes blog, and another 37% have access to CNCF website. More than 40% of the respondents explained cloud native technology through kubernetes blog, and 37% mentioned CNCF website.
We encourage CNCF members, project units and project ambassadors to provide supplier neutral content for CNCF blogs and share knowledge in a wider community. Knowledge of the community at large.
CNCF works with the Linux foundation to provide training and certification programs. In 2020, the project has been widely welcomed, with 17% kcsp, 16% KTP and 24% kubernetes from China. CNCF works with the Linux foundation to provide several training and certification programs. These are very popular in 2020, with 17% of kcsp, 16% of KTP and 24% of kubernetes certification companies coming from China.
In 2019, CNCF cooperated with Alibaba to launch free kubernetes and cloud native courses. At present, 20000 people have received the course training. CNCF, in collaboration with Alibaba, continues to offer free kubernetes and cloud native courses, which were launched in 2019 and have adopted more than 20000 people.
Investigation methods and objects
Thank you all for participating in this survey!
The survey was conducted in Chinese in December 2020 and January 2021, and 91% of the 439 respondents came from Asia. Last year, 300 respondents were surveyed. The survey was conducted in December 2020 and January 2021. Conducted in Putonghua, 91% of the 439 respondents came from Asia. Last year, we had 300 respondents.
About half (49%) of the members of the CNCF end user community are members of the report. Nearly half (49%) reported being part of the CNCF end-user community.
Most of the respondents came from enterprises. 40% came from units with more than 5000 employees, up from 23% last year. More than half come from units with more than 1000 employees, up from 40% last year. Most of the respondents came from enterprises. 40% are part of an organization with more than 5000 employees, up from 23% last year. More than half are part of organizations with more than 1000 employees, up from 40% last year.
More reading: Hacker News: Google’s contribution to all CNCF projects accounts for nearly 53%. China Mobile: cloud native white paper of telecom industry (with download) Amazon cloud technology self developed cloud native processor provides cost-effective computing power for graffiti intelligence. AWS Technology Summit 2018 Shanghai station is expected to have more than 50 technical forums and more than 6000 professionals attending, Top 10 cloud computing trends in 2021 Data information graph Forrester R esearch:2020 2010 global cloud computing scale reaches 241 billion US dollars IDC: in the first half of 2020, the overall market scale of China’s public cloud services (IAAs / PAAS / SaaS) reaches 8.4 billion US dollars People’s Bank of China: questionnaire survey report on urban depositors in the first quarter of 2019
If you want to get the full report, you can contact us by leaving us the comment. If you think the information here might be helpful to others, please actively share it. If you want others to see your attitude towards this report, please actively comment and discuss it. Please stay tuned to us, we will keep updating as much as possible to record future development trends.
RecordTrend.com is a website that focuses on future technologies, markets and user trends. We are responsible for collecting the latest research data, authority data, industry research and analysis reports. We are committed to becoming a data and report sharing platform for professionals and decision makers. We look forward to working with you to record the development trends of today’s economy, technology, industrial chain and business model.Welcome to follow, comment and bookmark us, and hope to share the future with you, and look forward to your success with our help.