MARCH 2022

NOVEMBER 2020
We help the world understand developers
We survey 30,000+ developers annually – across web, mobile, IoT, cloud, Machine Learning, AR/VR, games and
desktop – to help companies understand who developers are, what they buy and where they are going next.

WHO DEVELOPERS ARE WHAT THEY BUY WHERE THEY ARE GOING
Developer population sizing Why developers are adopting Emerging platforms – augmented &
Developer segmentation competitor products – and how you virtual reality, machine learning
can fix that
 T R U ST E D BY
the leading tech platforms

OUR CLIENTS
Can I share data from this report?
1. License Grant 2. Limitation of Liability
This report is licensed under the Creative Commons Attribution-NoDerivatives SlashDataTM , believes the statements contained in this publication to be based upon
Licence 4.0 (International) . Put simply, subject to the terms and conditions of this information that we consider reliable, but we do not represent that it is accurate or
license, you are free to: complete and it should not be relied upon as such. Opinions expressed are current
opinions as of the date appearing in this publication only and the information,
Share — You can reproduce the report or incorporate parts of the report into one or including the opinions contained herein, are subject to change without notice. Use of
more documents or publications, for commercial and non-commercial purposes. this publication by any third party for whatever purpose should not and does not
absolve such third party from using due diligence in verifying the publication’s
Under the following conditions: contents. SlashDataTM disclaims all implied warranties, including, without limitation,
warranties of merchantability or fitness for a particular purpose.
Attribution — You must give appropriate credit to SlashDataTM , and indicate if
changes were made. In that case, you may do so in any reasonable manner, but not in SlashDataTM , its affiliates, and representatives shall have no liability for any direct,
any way that suggests that SlashDataTM endorses you or your use. incidental, special, or consequential damages or lost profits, if any, suffered by any
third party as a result of decisions made, or not made, or actions taken, or not taken,
NoDerivatives — you cannot remix or transform the content of the report. You may based on this publication.
not distribute modified content.
The analyst of the developer economy | formerly known as VisionMobile
SlashData © Copyright 2022 | Some rights reserved

4
 Table of contents

Key insights | 6

Introduction | 7
A. Roadmap
B. Defining cloud native computing

1. The Cloud Native Landscape | 10

A. Market size
B. Regional usage of cloud native technologies
C. Where cloud native developers are running their code

2. Usage and Awareness of Kubernetes | 17

A. Kubernetes usage and awareness among backend developers
B. Profile of backend developers using Kubernetes
C. Does involvement in emerging areas differ according to Kubernetes usage?

3. Usage Trends of Orchestration and Serverless Tools | 25

A. Usage of orchestration tools
B. Usage of serverless tools

4. Security in the Cloud Native Space | 30

A. Impact on decisions

Appendix | 33
Methodology | 37

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 5
KEY INSIGHTS
• The global number of cloud native developers has grown by 1M in • The larger an organisation a backend developer works for, the
the last twelve months to 7.1M in Q3 2021. ➜ more proportionally likely they are to be using Kubernetes. ➜

• Overall, cloud native computing is now most widely adopted in • The top three vendors offering managed orchestration services —
North America (48%), South America (47%), and Oceania (50%). ➜ Amazon, Google, and Microsoft — dominate the market. ➜

• Cloud native developers are leveraging multiple deployment • The way in which developers engage with serverless products is
environments: 54% of cloud native developers use two or more changing. With developers choosing fewer serverless products to
environments to run their code, while only 40% of non-cloud native use simultaneously, competition among vendors in this landscape is
backend developers do. ➜ set to become fiercer. ➜

• As of Q3 2021, 30% of developers have used Kubernetes in the last • Among cloud native developers, those who are more security-
12 months — an estimated 5.1M backend developers. ➜ focussed are more likely to be able to affect tooling purchasing
decisions. ➜
• The most cited reason why backend developers don’t use
Kubernetes — given by 36% of those developers — remains a lack
of applicability or interest. ➜

6
Introduction
Introduction

A. Roadmap
The way software is developed has drastically changed since Kubernetes is arguably at the heart of cloud native applications and in
containers came about and cloud native technology gained popularity. the second chapter, we examine developers' usage and awareness of
Commissioned by Cloud Native Computing Foundation (CNCF), Kubernetes in an attempt to identify the drivers that facilitate its
SlashData has performed an in-depth analysis of the cloud native adoption across different developer profiles and technologies. We
developer ecosystem1 to better understand its current state and profile Kubernetes users in terms of professional status and the size of
expected evolution. the organisation these developers work.

The analysis is based on the 21st edition of SlashData’s Developer Chapter three gives an in-depth overview of various vendors’ container
Nation survey, which was fielded between June 2021 and August 2021 orchestration and serverless platform solutions and examines how the
and reached more than 19,300+ developers globally2. 3,900+ survey market share of these tools has changed over the last six months.
participants answered questions relating to the development of
backend services and the technologies they use. The report concludes by exploring security in the cloud native space.
Here, we focus on whether security is a modern priority, specifically
The first chapter of this report provides estimates for the global looking at the level of influence that security-concerned cloud native
number of cloud native developers and an in-depth regional overview developers have on tooling purchasing decisions.
of usage rates. Further, we deep-dive into the environments
developers are leveraging to deploy the code of their backend services
and we contrast cloud native with non-cloud native developer usage.

1 Cloud native developers are defined in the next section.

2 See the Methodology chapter for more information about how the survey was conducted.

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 8
Introduction

B. Defining cloud native computing

As in our previous report3, we use CNCF’s definition of cloud native This report focusses on developers who self-identify as working in the
computing as a guide for this analysis: backend services sector. These are developers who are writing server-
side code. The servers could be housed on-premise or in a third party's
“Cloud native technologies empower organisations to build and run data centre. Segmenting further, we focus on cloud native developers.
scalable applications in modern, dynamic environments such as public, In order to identify these developers, we examine what technologies
private, and hybrid clouds. Containers, service meshes, microservices, they are using to build backend services. Containers are the most
immutable infrastructure, and declarative APIs exemplify this approach. popular technology: three in five backend developers use containers.
These techniques enable loosely coupled systems that are resilient, We do not, however, class them all as cloud native developers. While
manageable, and observable. Combined with robust automation, they the use of containers may be an important first step in moving to cloud
allow engineers to make high impact changes frequently and native development, without automation, it remains exactly that: only
predictably with minimal toil.” the first step.

Returning to the CNCF definition, note that the key is the use of
automation to make high impact changes frequently and predictably
with minimal work. As CNCF was initially developed around
Kubernetes and container orchestration, we consider these to be at the
core of cloud native computing. Consequently, we have limited the
definition of cloud native developers to those that are using some sort
of container orchestration solution or serverless platform.

3 The State of Cloud Native Development Q1 2021

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 9
1. The Cloud Native
Landscape
1. The Cloud Native Landscape

A. Market size
Our estimates put the global number of cloud native developers for Usage of all cloud native native technologies shows an upward
Q3 2021 at 7.1M, or 41% of backend developers. The population population growth. This has been driven by a robust increase in the
includes 4.8M backend developers using container orchestration tools total number of backend developers, from 15.1M in Q3 2020 to 17.3M in
and 4.2M backend developers using serverless platforms, as well as a Q3 2021.
1.9M overlap of developers using both orchestration and serverless
technologies. For the cloud native developer population specifically, the population
has grown by 1M in the last 12 months. However, as a proportion of
To provide context, we only show the population trend of those using the backend developer population, the cloud native developer
containers, although not all of those developers are considered cloud community has remained stable — between 40% and 41% of the
native developers4. In Q3 2021, we estimate the global population of backend population.
developers using containers is 10.4M, or 60% of backend developers.
This is the most ubiquitous technology in backend services
development.

4 See the previous section for a definition of cloud native developers.

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 11
1. The Cloud Native Landscape

The cloud native developer ecosystem has added 1M in the past 12

months
Number of backend developers (in millions) (Q3 2020 n=3,978 | Q1 2021 n=3,600 | Q3 2021 n=3,941)

Growth of the backend developer population

Q3 2020

Q1 2021 10.4M
10.2M
Q3 2021
9.0M

7.1M
6.7M
6.1M

4.8M
4.6M
4.2M 4.1M
3.9M 3.8M

Cloud native Cloud functions or serverless architecture Container orchestration tools & Containers
management platforms

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 12
1. The Cloud Native Landscape

B. Regional usage of cloud native technologies

The usage of containers and cloud native technologies varies East Asia and the Greater China area show severe decline in usage of
significantly by region. Barring Eastern Europe and the Greater China cloud native technologies. In Q3 2020, four in ten backend developers
area — two regions which we will discuss shortly — the proportion of in East Asia were cloud native developers. 12 months later, the cloud
backend developers who use containers and container orchestration native developer population has dropped by 8 percentage points, to
tools has remained relatively unchanged in the past year — between 32%. Usage of all cloud native technologies have simultaneously fallen.
Q3 2020 and Q3 2021. The adoption of serverless solutions has shown A recent CNCF report, showed that developers in China thought
more variability — increasing in North America, South America, and complexity, security, and reliability were the top challenges with using
Oceania and decreasing in Western Europe, East Asia, and the Greater containers.
China area within the last 12 months.

Overall, cloud native computing is now most widely adopted in North

America (48%), South America (47%), and Oceania (50%).

With 44% of backend developers classed as cloud native, Eastern

Europe is the fourth region where cloud native computing is most
widely adopted. The region has shown a significant increase of 10
percentage points in the cloud native developer population within the
last 12 months. It is also the region which has had the most significant
redress in terms of gender dynamics: in Q3 2020, 95% of cloud native
developers in Eastern Europe were male; the highest proportion of
males in any region. By Q3 2021, this has decreased to 91% — almost
on par with the global cloud native population which is 90% male, 8%
female, and 2% gender fluid/multiple genders/non-binary,
agender/none, or preferred not to specify.

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 13
1. The Cloud Native Landscape

East Asia shows decline in usage of cloud native technologies

% of backend developers in each region (Q3 2020 n=3,978 | Q3 2021 n=3,941)

Usage of cloud native technologies across regions

Q3 2020 Q3 2021
Cloud native

48% 47% 51% 47% 50%

44% 43% 45% 40% 39% 44% 43%
32% 36% 34% 32% 36% 36%
architecture
Serverless

45%
28% 31% 33% 27% 33% 35%
24% 22% 25% 25% 22% 20% 26% 22%
14% 15% 12%
orchestration
Containers

37% 42%
31% 30% 30% 30% 26% 29% 29% 31% 25% 29%
23% 19% 23% 23% 19% 23%

70% 79%
70%
Containers

70% 68% 65% 69% 66%

60% 61% 55% 57% 59%
45% 52% 50% 51%
43%

Western Europe & North America East Asia ex. South Asia Eastern Europe, Middle East & South America Greater China area Oceania
Israel Greater China Russia & Former Africa
CIS

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 14
1. The Cloud Native Landscape

C. Where cloud native developers are running their code

In our survey, we asked backend developers where their code runs, Comparing cloud native developers to their non-cloud native peers,
with potential answers including on a public cloud, private cloud, cloud native developers are more likely than other backend
hybrid cloud, multi-cloud, and on-premise servers. These options are developers to use cloud environments, illustrating how cloud native
not mutually exclusive. For greater clarity, within the survey, we define development is all about leveraging the capabilities of the cloud. On-
a private cloud as a cloud that is only available to certain users premise servers, together with mainframe, are the only two
regardless of whether it is hosted on-premise or in a third party's data environments where non-cloud native developer use exceeds cloud
centre. We also define hybrid clouds as using a combination of public native developer use.
and private clouds for a single project and multi-clouds as using
multiple public clouds for a single project. Over the last six months — from Q1 2021 to Q3 2021 — usage of almost
all environments among cloud native developers has decreased
Cloud native developers are leveraging multiple deployment slightly. On-premise servers are the exception: usage rose 4
environments. 54% of cloud native developers use two or more percentage points in this period. Cloud native developers seem
environments to run their code, while only 40% of non-cloud native unwilling to abandon existing on-premise servers — often using this
backend developers do. The flexibility of cloud native development environment in addition to cloud environments. This behaviour could
enables organisations to operate with distributed environments and possibly be due to company policies requesting that developers run
allocate workloads to compute resources best suited for any particular their code on on-premise servers to maintain a certain level of security
job. and control over the data. As we’ll see in chapter four, security-
focussed cloud native developers are more likely than those who aren’t
Analysing which environments cloud native developers are using in security-focussed to be company decision makers.
order to take advantage of this flexibility yields public cloud, on-
premise servers, and private cloud as the three most popular
environments — all used by more than two in five cloud native
developers.

Note: In Appendix B, we deep-dive to examine where cloud developers involved in telecommunications run their code.

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 15
1. The Cloud Native Landscape

Cloud native developers rely more heavily on cloud environments

than their non-cloud native peers
% of backend developers who use each environment (Q1 2021 n=1,798 | Q3 2021 n=2,039)

Where backend developers run their code

50% Cloud native developers

48% 47%
40%
Non-cloud native developers
Q1 2021

34%
32%
26%
20%
16%
11%
9% 7%

52% 50%
43% 41%
Q3 2021

38%
32%
29%
22%
15% 13%
10% 6%
Public cloud On-premise servers Private cloud (cloud only Hybrid cloud (using public Multi-cloud (using multiple Mainframe
available to certain users) and private clouds for a public clouds for a single
single project) project)

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 16
2. Usage and
Awareness of
Kubernetes
2. Usage and Awareness of Kubernetes

A. Kubernetes usage and awareness among backend developers

We asked more than 5,500 backend developers about their awareness In the rest of this chapter, we deep-dive into the profile of developers
and usage of both Kubernetes and containers. This large sample who have endorsed Kubernetes — in the process, understanding who
comprises both cloud native and non-cloud native developers and hasn’t adopted this technology.
provides a clear picture of how well-known these cloud native
technologies are among the whole population of backend developers. One thing is already clear: Kubernetes appeals particularly to backend
developers who are already interested in or heading towards cloud
This section explores backend developers’ awareness and usage of native computing. Among cloud native developers, 47% use
Kubernetes and containers. As of Q3 2021, 61% of backend developers Kubernetes and only 3% have never heard of this technology.
had used containers in the last 12 months and 30% of developers used
Kubernetes to orchestrate these: an estimated 5.1M backend In Q3 2021, we also began tracking usage and awareness of another
developers. The overall usage of Kubernetes is reasonably stable — cloud native technology, microservices. 48% of backend developers
increasing slightly by 2 percentage points between Q3 2020 and Q3 have used microservices in the last 12 months. This increases to 66%
2021. among cloud native developers.

Although the percentage of developers using Kubernetes has

remained relatively stable, many backend developers remain unsure of
what it can do for them. One in ten developers have never heard of
Kubernetes, and roughly one in four are unsure what it does. However,
the most cited reason why backend developers don’t use Kubernetes
— given by 36% of these developers — remains lack of applicability
or interest. Therefore, it may be that most backend services are still
simple enough to require container orchestration or that some of those
developers rely on Container as as Service (CaaS) solutions without
knowledge of or interest in the orchestration engines that power these.

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 18
2. Usage and Awareness of Kubernetes

Nine in ten developers are aware of Kubernetes

% of backend developers (Q3 2020 n=5,543 | Q3 2021 n=5,143)

Cloud native technology usage and awareness

Q3 2020 Q3 2021
Kubernetes

35% 36%
28% 30%
21% 22%
15%
9%

61%
56%
Containers

23% 22% 7% 4%
13% 12%
Microservices

48%

26%
17% 6%

Used it in the past 12 months Heard of it but not sure what it Aware of it, but not Never heard of it
does interested/does not apply

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 19
2. Usage and Awareness of Kubernetes

B. Profile of backend developers using Kubernetes

What do the backend developers who use Kubernetes look like? At the In terms of education, backend developers who use Kubernetes are
overview level, we define two categories and make a distinction more likely than those who don’t use this technology to be formally
between professionals and non-professionals5. Kubernetes users are educated: 49% learnt to code by taking an undergraduate degree in
professionals to a large extent: nine in ten of these developers identify computing or software engineering and 27% by taking a postgraduate
as such. Backend developers who don’t use Kubernetes are less likely degree. In comparison, among backend developers who don’t use
to be professionals: only 73% are. In general, the backend services Kubernetes, 43% learnt to code through an undergraduate degree in
sector has the highest concentration of professionals (77%) across all computing and 17% through a postgraduate degree. In addition,
sectors we track in our surveys. around seven in ten of all backend developers — regardless of whether
they use Kubernetes or not — learn to code through self-study.
Further, we explore the size of organisations that backend developers
who use Kubernetes work for — again contrasting with the backend We conclude that Kubernetes developers are more likely to be
developers who don’t use Kubernetes, to gain context. Organisation professionals and to be working for very large enterprises than
size categories are given in Appendix A. As a general trend, the larger backend developers who do not use Kubernetes. Heading to the
an organisation a backend developer works for, the more microscope briefly, we see that they are also security-focussed: one in
proportionally likely they are to be using Kubernetes. Close to one in four backend developers who use Kubernetes have recently chosen
four developers working for very large enterprises use Kubernetes. The to migrate their apps to the cloud due to security reasons. However,
adoption of Kubernetes arguably requires a significant investment of backend developers who do not use Kubernetes are even more likely
time and resources with the scale and complexity of projects in larger to migrate due to security; 29% have chosen to do so. We deep dive
organisations driving the use of container engines. Larger further into the issue of security in the cloud native space in chapter
organisations have better resources and use cases to adopt four.
Kubernetes.

5If a developer is involved professionally in building backend services, we count them as a professional, regardless of whether they are also a hobbyist or student in that same sector.
Non-professionals could be hobbyists, students, or both.

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 20
2. Usage and Awareness of Kubernetes

9 in 10 Kubernetes developers are professionals

% of backend developers (Q3 2021 n=7,090)

Mix of professionals and non-professionals in the Kubernetes developer community

2%
4%
5% 10%
4% non-professional 27%
12% non-professional
Hobbyists and students
17%
10%

Exclusively students 14%

Exclusively hobbyists
90% 73%
professional professional
72%
Professionals, and hobbyists or students
59%

Exclusively professionals

Backend developers using Kubernetes Backend developers not using Kubernetes

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 21
2. Usage and Awareness of Kubernetes

The larger an organisation a backend developer works for, the

more proportionally likely they are to be using Kubernetes
% of backend developers by the size of the organisation they work for (Q3 2021 n=5,934)

How organisation size influences Kubernetes usage

40%

30%
Backend developers
who use Kubernetes

20%

10%
Backend developers who
do not use Kubernetes

0%
Freelancer Small business Mid-market Small enterprise Large enterprise Very large
enterprise
Organisation size

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 22
2. Usage and Awareness of Kubernetes

C. Does involvement in emerging areas differ according to

Kubernetes usage?
In this section, we examine which emerging areas backend developers Emerging areas with the largest differential in involvement between
using Kubernetes are currently working in. These areas include: mini backend developers who use Kubernetes and those who don’t are fog
apps, robotics, 5G, quantum computing, fog and edge computing, and edge computing, biometrics for ID verification, and blockchain
cryptocurrencies, and other blockchain applications, among others. applications. Developers who use Kubernetes are 9 to 10 percentage
points more likely than those who don’t to be working in these areas.
Mini apps, computer vision, blockchain applications, The usage of Kubernetes in these areas is also growing: biometrics for
cryptocurrencies, and biometrics for ID verification are the areas ID verification and blockchain applications both grew eleven
with the largest proportion of backend developers using Kubernetes percentage points in popularity among Kubernetes users over the last
— more than one in five developers in each of these areas uses twelve months, while fog and edge computing grew 3 percentage
Kubernetes. points6.

Backend developers who use Kubernetes are involved in a higher The emerging area with the least differential in involvement is robotics.
number of emerging areas than their peers who don’t use this Whether backend developers use Kubernetes has little effect on their
technology: 48% of Kubernetes users work in two or more emerging rate of engagement in this area, which sits at 12%7.
areas, while backend developers who don’t use Kubernetes are 10
percentage points less likely to work in more than one area.

6Our previous report, The State of Cloud Native Development Q1 2021, offered significant insights into edge computing in the cloud native space.
7There are certain challenges with running the Robot Operating System (ROS) on Kubernetes, including Kubernetes’ lack of support for dynamic ports. However, recent work with
ROS 2 has proved promising in this area, and as such, this is an interesting emerging area for Kubernetes growth in the future.

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 23
2. Usage and Awareness of Kubernetes

Backend developers who use Kubernetes are involved in a higher

number of emerging areas than their peers who don’t use this
technology
% of backend developers (Q3 2021 n=1,301) The influence of Kubernetes usage on emerging areas
Mini apps 37%
35%
Computer vision 25%
24%
Blockchain applications other than cryptocurrency 23%
14%
Cryptocurrencies (e.g. Bitcoin) 22%
18%
Biometrics for ID verification (e.g. fingerprints/iris/voice verification) 22%
13%
Conversational platforms/voice search 20%
14%
Emerging areas

Fog/edge computing 17%

7%
5G 15%
10%
12%
Robotics
12%
DNA computing/storage 9%
5%
8%
Drones Backend developers who use
8%
8% Kubernetes
Quantum computing 6%
Brain/body computer interfaces (e.g. advanced prosthetics, neural lace) 8%
5% Backend developers who do not use
Haptic feedback 8% Kubernetes
6%
Self-driving cars 6%
5%
Hearables 6%
4%

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 24
3. Usage Trends of
Orchestration and
Serverless Tools
3. Usage Trends of Orchestration and Serverless Tools

A. Usage of orchestration tools

While Kubernetes has become the industry standard, developers are Some developers are finding the possibility of using a self-managed
leveraging it in different ways. Developers might self-host Kubernetes solution in conjunction with a vendor’s offerings particularly
clusters while others use container orchestration services to help appealing. 42% of developers using Microsoft Azure Service Fabric,
manage their Kubernetes clusters. The majority of container 60% of developers using Red Hat Advanced Cluster, and 44% of
orchestration tools — including five of the top six most used — are developers using IBM Cloud Kubernetes Service also build their own
managed Kubernetes solutions. The market is strongly shaped by container management solution or self host.
Kubernetes and managed Kubernetes services.
Further, the proportion of developers choosing a single orchestration
Overall, the top three vendors offering managed orchestration tool has decreased: 49% of backend developers who use or are aware
services — Amazon, Google, and Microsoft — dominate the market. of any orchestration tool used only one tool in Q1 2021. Six months
24% of backend developers who use or are aware of any orchestration later, 46% of these developers use only one tool. More developers are
tool are currently using Amazon’s Elastic Container Service (ECS), choosing to use multiple orchestration solutions in order to leverage
making it the most widely adopted solution. After a 3 percentage point different advantages from each. Briefly scanning the satisfaction
decline in the last six months, Google Kubernetes Engine (GKE) is in scores of these vendor products, no single solution dominates across
second place, with 21% of developers using it. the attributes important to backend developers who use orchestration
tools: pricing, documentation and sample code, technical support,
Interestingly, 30% of developers using orchestration tools are building integration with other systems, suitability, and feature set, as well as
their own container management solution or they are self-hosting an ease and speed of development.
existing offering. In the last six months, this proportion has grown by 3
percentage points. These developers are circumventing vendor
offerings and might instead be turning towards standalone
orchestration engines, such as Kubernetes. Self-managed solutions
give developers full control, allowing them to tailor functionalities to
their specific needs which may be required for custom systems.

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 26
3. Usage Trends of Orchestration and Serverless Tools

Amazon ECS leads the container orchestration market

% of backend developers who use or are aware of any orchestration tool (Q1 2021 n=968 | Q3 2021 n=1,081)

Usage of orchestration tools over the last six months

Q3 2021 Q1 2021

Amazon Elastic Container Service 24%

25%

Google Kubernetes Engine 21%

24%

Amazon Elastic Kubernetes Service 19%

21%

Microsoft Azure Kubernetes Service 18%

17%

Red Hat OpenShift 10%

11%

DigitalOcean Kubernetes 8%
9% Managed Kubernetes solutions

Microsoft Azure Service Fabric 9%

9%

Red Hat Advanced Cluster 5%

6%

IBM Cloud Kubernetes Service 6%

6%

VMware Tanzu Kubernetes Grid 6%

6%
4%
Linode Kubernetes Engine
5% Managed Kubernetes solutions
We develop this in-house/self- 30%
hosted 27%

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 27
3. Usage Trends of Orchestration and Serverless Tools

B. Usage of serverless tools

Serverless architecture is lauded for its self-managed, detailed scaling, quite surprising to see such a dramatic decrease in usage of 6
as well as its automatic resource provisioning. The key benefit is seen percentage points from Q1 2021 to Q3 2021. This decrease might
as its ability to scale, which is transparent and managed to the level of partly be driven by a decrease in usage of GKE by 3 percentage points
individual requests. This in turn reduces costs, as developers pay only over the same period8.
for the resources they use and there is no need to pay for idle servers.
Here, we investigate the landscape of serverless product offerings. Although less extreme, both AWS Lambda and Google Cloud
Functions decreased in usage by 3 percentage points over this six
AWS Lambda is by far the most widely adopted product, with 43% of month period. Azure Functions alone remains steady in this trend-
backend developers who use serverless architecture choosing to use it. decreasing landscape, maintaining 26% usage.
Google’s serverless products occupy second place with 29% usage for
Google Cloud Functions, and fourth place with 15% usage of Google It seems that the way in which developers engage with serverless
Cloud Run. products is changing. In the past, developers would leverage multiple
serverless products simultaneously. In Q1 2021, 23% of backend
Google Cloud Functions is the established serverless offering from developers who used or were aware of any serverless product used
Google. As an example of a Function as a Service (FaaS) product, it three or more serverless products. Six months later, only 19% of
offers the typical event-driven function-triggering framework. developers do. With developers choosing to use fewer products,
However, there are only a small set of languages in which one can competition in this landscape is set to become fiercer.
write code for Google Cloud Functions — JavaScript/Node.js, Python
3, or Go — and it places restrictive constraints on how code is Vendors’ hold on serverless architecture is not ubiquitous: 13% of
deployed. As an alternative, Google Cloud Run is the vendor’s latest developers who use or are aware of any serverless product choose to
serverless offering. A type of serverless container, Google Cloud Run host in their own data centres.
affords impressive flexibility and works cohesively with Google’s
orchestration tool, GKE. Google Cloud Run has been rising in
popularity over the couple of years since its launch — as such, it is

8 Usage of orchestration tools is discussed in more detail in the previous section.

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 28
3. Usage Trends of Orchestration and Serverless Tools

Usage of Google Cloud Run dropped by 6 percentage points in

the past six months
% of backend developers who use or are aware of any serverless product (Q1 2021 n=911 | Q3 2021 n=951)
Usage of serverless tools over the last six months
Q3 2021 Q1 2021
43%
AWS Lambda
46%
Google Cloud Functions 29%
32%
26%
Azure Functions 26%
15%
Google Cloud Run
21%
11%
Netlify Functions
10%
10%
Cloudflare Workers
10%
6%
Twilio Functions
9%
6%
IBM Cloud Functions (OpenWhisk)
8%
4%
Oracle Functions (Fn) 6%
3%
Alibaba Cloud Function Compute 6%
2%
Yandex Cloud Functions
4%
3%
PubNub Functions/BLOCKS
4%
4%
Knative
3%
13%
We host this in our own data centre
16%

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 29
4. Security in the
Cloud Native Space
4. Security in the Cloud Native Space

A. Impact on decisions
There are many tiers of cloud native developers: these range from In agreement with this analysis: 10% of security-focussed cloud native
those high up in a business hierarchy that make or influence developers are not involved in purchasing decisions, while this rises to
purchasing decisions, down to developers who are not involved at all 14% among cloud native developers in general.
in these selections. Here, we show the breakdown of cloud native
developers according to economic decision-making power within an Looking at security in the cloud native landscape from another angle:
organisation and how this differs depending on attitudes towards security issues are relatively unlikely to affect adoption of serverless
security. Those who are more security-focussed — defined as tools — only 6% of cloud native developers would adopt a serverless
developers who have recently chosen to migrate their apps to the tool due to security, ranking this consideration 15th out of 17. Similarly,
cloud due to security reasons — are more likely to be able to affect 5% of cloud native developers would reject a serverless tool due to
tooling purchasing decisions. security concerns.

32% of security-focussed cloud native developers make the final

selection decision for team or company tools. This is 10 percentage
points higher than in the general cloud native population. Security-
focussed cloud native developers are also significantly more likely to
be responsible for specifications, able to approve expenses on tools
and the overall team budget for developer tools.

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 31
4. Security in the Cloud Native Space

Security-focussed cloud native developers are more likely to be

economic decision-makers
% of cloud native developers that have each level of influence (Q3 2021 n=1,828)

Cloud native developers’ influence on tooling purchasing decisions

7%
We are not buying any tools or components
7%

10%
Not involved in selection/purchase decisions
14%

49%
Buying as an individual, for my own use
45%

50%
Making recommendations or influencing decision makers
51%

38%
Responsible for specifications
29%

32%
Making the final selection decision for team/company tools
22%

22% Security-focussed cloud native

Approving expenses on tools & components
13% developers

18% Cloud native developers

Approving the overall team budget for developer tools
10%

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 32
Appendix

33
Appendix A

We asked professional backend developers what size organisation they work for. Defining:
• Freelancer developers as those who work for themselves;
• Small business developers as those who work for organisations with between two and 50 employees;
• Mid-market businesses between 51 and 500 employees;
• Small enterprises between 501 and 1,000 employees;
• Large enterprises between 1,000 and 5,000 employees;
• Very large enterprises with more than 5,000 employees.

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 34
Appendix B

Telecommunications backend developers are 7 percentage points

more likely to be cloud native developers than backend
developers in general
% of backend developers (Q3 2021 n=3,941)

Usage of cloud native technologies

Backend developers involved in telecommunications Backend developers

67%

60%

48%

41%
38%

28%
24% 25%

Cloud Native Developer Cloud functions or serverless architecture Container orchestration tools & Containers
management platforms

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 35
Appendix B

Cloud native developers working in the telecommunications

industry leverage a higher number of deployment environments
than their non-cloud native peers
% of backend developers involved in telecommunications who use each environment (Q3 2021 n=198)

Where cloud native developers involved in telecommunications run their code

Cloud native developer involved in telecommunications Non-cloud native developers involved in telecommunications

64%
57%
53%

42%
35% 35%
31%
24%
19%
14%
10%
8%

On-premise servers Private cloud (cloud only Public cloud Hybrid cloud (using public Multi-cloud (using multiple Mainframe
available to certain users) and private clouds for a public clouds for a single
single project) project)

THE STATE OF CLOUD NATIVE DEVELOPMENT Q3 2021 | KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION 36
METHODOLOGY
The Developer Economics Survey
Developer Economics 20th edition reached 19,000+ respondents from 155 To eliminate the effect of regional sampling biases, we weighted the
countries around the world. As such, the Developer Economics series regional distribution across eight regions by a factor that was determined
continues to be the most global independent research on mobile, desktop, by the regional distribution and growth trends identified in our Developer
industrial IoT, consumer electronics, embedded, third party app Economy research. Each of the separate branches: mobile, desktop,
ecosystems, cloud, web, game, AR/VR, and machine learning developers iIndustrial IoT, consumer electronics, embedded software, third party app
and data scientists combined, ever conducted. The report is based on a ecosystems, cloud, web, games, augmented and virtual reality, and data
large-scale online developer survey designed, produced, and carried out science and machine learning were weighted independently and then
by SlashData over a period of ten weeks between November 2020 and combined.
February 2021.
To minimise other important sampling biases across our outreach
Our respondents came from a broad age spectrum, from young coders channels, we weighted the responses to derive a representative
who are under 18 to the seasoned ones over 55. As software development distribution for technologies used and developer segments. Using
is still a man’s world, 79% of our respondents were male and 20% female, ensemble modelling methods, we derived a weighted distribution based
excluding other options and those who did not specify their gender. on data from independent, representative channels, excluding the channels
of our research partners to eliminate sampling bias due to respondents
Respondents were asked which types of projects they are involved in out
who were recruited via these channels.
of the 13 under study, namely web apps / SaaS, mobile apps, desktop
apps, backend services, augmented reality, virtual reality, games, data Again, this was performed separately for each of mobile, industrial IoT,
science, machine learning / artificial intelligence, industrial IoT, consumer consumer electronics, embedded software, third party app ecosystems,
electronics devices, embedded software, and apps/extensions for third desktop, cloud, web, games, augmented and virtual reality, and data
party ecosystems. They also told us if they are into their areas of science and machine learning.
involvement as professionals, hobbyists, or students - or as any
combination of these - and how many years of experience they have in
For more information on our methodology please visit
each.
https://www.slashdata.co/methodology.

37
We help you understand Developers.
If you could speak to 30,000+ developers what would you ask them?

Alexes Mes
Mathematical and Statistical Consultant
alexes@slashdata.co

SlashData Ltd.
19-21 Hatton Gardens
London, EC1N 8BA
Konstantinos Korakitis United Kingdom
Research Operation Manager +44 845 003 8742
konstantinos@slashdata.co hello@slashdata.co

38