Comparative interpretation of "CNCF China Cloud Native Survey Report 2020"

Introduction to On April 28, CNCF released the 2020 China Cloud Native Survey Report. This is the fourth cloud native survey conducted by CNCF in China. On the whole, it represents cloud native containers, serverless and service mesh. Both in terms of the user groups and the scale of implementation, they are in a stage of rapid growth.

On April 28, CNCF released the 2020 China Cloud Native Survey Report. This is the fourth cloud native survey conducted by CNCF in China. The purpose is to help relevant practitioners have a deeper understanding of the progress of cloud native in China.

report address: https://www.cncf.io/blog/2021/04/28/cncf-cloud-native-survey-china-2020/

The author participated in this questionnaire survey in December last year. From the perspective of the content of the questionnaire, the two reports in 2020 and 2019 have relatively good continuity. The report focuses on the changes in the scale of cloud native landing in China, as well as the changes in the use and distribution of various open source projects or products.

The questionnaire sample is highly representative. In this year's report, a total of 439 people participated in the survey, and 49% were members of the CNCF's end user community. Among them, the scale of enterprises is greater than 5000 people/100-499 people/1000-4999 people/500-999 people, accounting for 40%/19%/12%/11% respectively, and other scale groups do not exceed 10%; the industry distribution is based on Software/technology is the main industry, more than 60%, telecommunications/financial services/education are also industries with more practice, between 5%-10%; the respondents are mainly software architects/back-end engineers/DevOps engineers, total More than 90% (not deduplicated). These characteristics also seem to be more compatible with the companies around us that are practicing cloud native.

Next, let's interpret some of the main conclusions related to Serverless in the report:

Serverless related:

• Serverless is continuing to grow, 31% of units use serverless in production, 41% are evaluating, and 12% plan to use it in the next 12 months.

31% are in use and 41% are being evaluated, which is close to the usage level of containers in 2019 (49% are in use and 32% are being evaluated). The value of serverless has been recognized by most companies, and the serverless programming model is also helping developers to integrate into the business in a brand new posture. Take Alibaba as an example. Serverless has the momentum of microservices that year. Taobao, AutoNavi, Alipay, Xianyu, Fliggy, B2B, almost all businesses are applying Serverless technology to production scenarios, and also open source Serverless Devs, Midway and many other projects, I believe that in the near future, there will be star open source projects such as Dubbo and Nacos in the microservice field.

• Among users who use serverless, 45% use hosted platforms and 20% use installable software.

The hosting platform is still the first choice for users to implement Serverless technology, and the four hosting platforms are all products in the form of FaaS. Take the number one Alibaba Cloud Function Computing as an example. It is an event-driven fully managed serverless computing service that provides a function-level programming paradigm. Different from the application-level serverless form, responding with functions to the business has a larger transformation cost to the architecture. Choosing a hosting platform has a higher cost performance. Although 20% of users chose installable software in this report, more than 30% of them are from Knative. As we all know, Knative is not a FaaS model, but a best practice of CaaS. If you exclude Knative and choose Kubeless/Open Whisk to build a function management platform, there are very few users. It is worth noting that the usage of Kubeless has dropped from 29% last year to 11% this year.

• Alibaba Cloud Function Computing is still the most popular product in China, and it is favored by 35% of users.

This is not unrelated to the performance and maturity construction of Alibaba Cloud's underlying infrastructure such as system software, computing, network, storage, and containers, which makes functional computing unique in its runtime capabilities and stability. In addition, the key to the selection of FaaS products is not only the product's own capabilities, but also the integrity of the entire product ecosystem, such as the type and maturity of triggers related to FaaS, observable capabilities on the cloud, and various standards of open source Compatibility, these all depend on the overall product capabilities of cloud vendors. The overall product richness and product capabilities on Alibaba Cloud have lowered the threshold for users to switch to functions to manage computing resources.

In the first domestic "Cloud Native User Survey Report" released by the Academy of Information and Communications Technology in October last year, Alibaba Cloud accounted for 66% of domestic serverless users; in March this year, Forrester released the FaaS platform evaluation report for the first quarter of 2021. Alibaba Cloud's functional computing product capabilities rank first in the world and enter the FaaS leader quadrant, the only one in China. These also confirmed the basic logic of customers' selection in the FaaS field.

In addition, this year's China Cloud Native Survey Report also gave some main conclusions related to containers, Service Mesh, and open source tools.

Container related:

• Containers continue to grow rapidly. 68% of organizations use containers in the production process, an increase of 39% compared to last year, and an increase of 240% compared to two years ago.
• Kubernetes has become ubiquitous, and the proportion of Kubernetes used in production has increased from 72% last year to 82%.
• Public cloud has returned to the top of the list, with 46% of survey respondents using public cloud, which is higher than 36% last year but lower than 51% in 2018.

Service Mesh related:

• Service grid is very popular, 38% of units use service grid in production, but its complexity still needs to be reduced. 42% of respondents are evaluating service grids, and 11% plan to use them in the next 12 months. Therefore, it is expected that service grids will become a growth area in the coming years.
• Istio, Consul, and Linkerd are still the three most popular solutions in the service mesh field.

Open source tools related:

• The degree of automation continues to increase. 30% of survey respondents use automated release cycles in production, and 73% run CI/CD pipelines.
• Observability tools are widely used. 95% of organizations use monitoring tools, 94% use logs, and 85% use distributed tracking.

On the whole, the containers, serverless, and service mesh that represent cloud native are in a rapid growth stage, both in terms of user groups and scale of landing. The community has also become more prosperous, and related open source tools are helping developers to apply these three types of technologies to production more efficiently and stably. At present, China is CNCF's third largest contributor base (contributors and committers) after the United States and Germany. Among them, Alibaba has entered the CNCF sandbox for 6 open source projects (a total of 14 in China), namely Dragonfly, OpenKruise, OpenYurt, Vitess, ChaosBlade and Fluid.

Finally, we compare the data in the previous two reports together. Through the changes in the data, we may be able to provide some selection references for enterprises that plan or are landing cloud native. The following is the detailed comparison data.

		2019 Nian	2020 Nian
questionnaire sample	Published	2020 Nian 10 Yue	2021 Nian 4 Yue
	respondents group amount	300 people, 97% are from Asia	439 people, 91% are from Asia
	size of the company where the respondent is	>500-999: 2.3.3% 499: 2.3.3% 100-500-999: 2.3.3% -99: 9.0% 10-49: 13.0%>10: 3.7%	>5000 people: 40% 1000-4999: 12% 500-999: 11% 100-499: 19% 50-99: 6% 10- 49:. 9%> 10: 2%
container	ratio used in the production of containers	49%	68%
	ratio used in the production of K8s	72%	82%
	public cloud usage	36%	46 is%
	K8S production cluster size	44% cluster 50:: 13 or more clusters 6-100 cluster 5000%: 10%	more clusters 6-100: 47% cluster 50: cluster 21% 5000: 15%
	biggest challenge to the use or deployment vessel	use / deployment complexity: 53% Security: 39% Insufficient training: 36% Network: 36% Reliability and monitoring: 35%	Usage/deployment complexity: 58% CI/CD: 44% Monitoring: 38% Security: 36+20 years report does not give a hundred percent, the above report data from the histogram calculation Quasi
	a method of packaging Kubernetes	64%	54 is%
	service agent	NGINX: 54 is HAProxy%: 18 is% F5: 16% Envoy: 15%	: 39% Envoy: 29% HAProxy: less than 20% F5: less than 15%
	Kubernetes environment	Docker Kubernetes: 42 is Minikube%: cloud provider management Kubernetes: 14%	Docker Kubernetes: 37 [% cloud provider management Kubernetes: 28% Minikube:% less than 21 is
	work Load auto-scaling	Do not want to auto-scaling any work: 28% scheduled auto-scaling Stateless applications: 54% scheduled auto-scaling tasks/queue processing: 26% scheduled auto-scaling State applications: 23%	do not want to auto-scaling any work: 13% plan to automatically scale stateless applications: 69% plan to scale tasks/queue processing: 35% plan to scale automatically State applications: 34%
Serverless	Serverless deployment method	Installable software hosting platform: 36% 22%	Hosting platform: 45% Installable software: 20%
	Serverless hosting platform usage distribution	Cloud and Tencent Function Computing FC: 46% AWS Lambda: 34% 12% Azure Functions: less than 10%	Cloud Function Compute FC: 35% AWS Lambda: 19% Tencent Cloud SCF: less than 10% Huawei Cloud FS: less than 10% (the share of the 4 manufacturers has decreased For the reason, I consulted the CNCF official: In this year’s accounting algorithm, the denominator contains <span>Installable Serverle SS) </ span>
	Serverless install software usage profile	Kubeless: 29% Knative: 22 is% the Apache OpenWhisk: 20 is% the Virtual Kubelet: exceeds 10%	Knative: 31 is% Kubeless:. 11% the Virtual Kubelet : 9% Apache OpenWhisk:% less. 9
-Service and the Mesh storage	-Service the Mesh	Istio: exceeds 40% Consul: exceeds 20% Linkerd:% exceeds 10	Istio: nearly 40% Consul: more than 30 % Linkerd: exceeds 30%
	cloud native storage	the Ceph: 24% Amazon EBS: 23 is% of the container storage interface the CSI: 18 is%	the Ceph: 31 is% the CSI: 26 is% the Gluster: 15% using
The proportion of other open source projects used	Prometheus	57%	72%
	CoreDNS	35%	64%
	ETCD	37 [%	77%
	Harbor	27%	47%
	the CNI	22 is%	53 is%
	GRPC	29%	47%

> Copyright Statement: content of this article is contributed spontaneously by Alibaba Cloud real-name registered users. The copyright belongs to the original author. The Alibaba Cloud Developer Community does not own its copyright and does not assume corresponding legal responsibilities. For specific rules, please refer to the "Alibaba Cloud Developer Community User Service Agreement" and the "Alibaba Cloud Developer Community Intellectual Property Protection Guidelines". If you find suspected plagiarism in this community, fill in the infringement complaint form to report it. Once verified, the community will immediately delete the suspected infringing content.