NOVEMBER 2020

Transcription

NOVEMBER 2020ED RHS ELI MBB E1PU EC 202D

We help the world understand developersWe survey 30,000 developers annually – across web, mobile, IoT, cloud, Machine Learning, AR/VR, games anddesktop – to help companies understand who developers are, what they buy and where they are going next.WHO DEVELOPERS AREWHAT THEY BUYWHERE THEY ARE GOINGDeveloper population sizingDeveloper segmentationWhy developers are adoptingcompetitor products – and how youcan fix thatEmerging platforms – augmented &virtual reality, machine learning

T R U ST E D BYthe leading tech platformsOURCLIENTS

Can I share data from this report?1. License Grant2. Limitation of LiabilityThis report is licensed under the Creative Commons Attribution-NoDerivativesLicence 4.0 (International) . Put simply, subject to the terms and conditions of thislicense, you are free to:SlashDataTM , believes the statements contained in this publication to be based uponinformation that we consider reliable, but we do not represent that it is accurate orcomplete and it should not be relied upon as such. Opinions expressed are currentopinions as of the date appearing in this publication only and the information,including the opinions contained herein, are subject to change without notice. Use ofthis publication by any third party for whatever purpose should not and does notabsolve such third party from using due diligence in verifying the publication’scontents. SlashDataTM disclaims all implied warranties, including, without limitation,warranties of merchantability or fitness for a particular purpose.Share — You can reproduce the report or incorporate parts of the report into one ormore documents or publications, for commercial and non-commercial purposes.Under the following conditions:Attribution — You must give appropriate credit to SlashDataTM , and indicate ifchanges were made. In that case, you may do so in any reasonable manner, but not inany way that suggests that SlashDataTM endorses you or your use.NoDerivatives — you cannot remix or transform the content of the report. You maynot distribute modified content.SlashDataTM , its affiliates, and representatives shall have no liability for any direct,incidental, special, or consequential damages or lost profits, if any, suffered by anythird party as a result of decisions made, or not made, or actions taken, or not taken,based on this publication.The analyst of the developer economy formerly known as VisionMobileSlashData Copyright 2021 Some rights reserved4

Table of contentsKEY INSIGHTS1. INTRODUCTIONA.B.C.D.BackgroundRoadmapMarket sizeRegional usage of cloud native technologies2. USAGE AND AWARENESS OF KUBERNETESA. Kubernetes and containers usage and awareness among backend developersB. Profile of backend developers using Kubernetes3. EDGE COMPUTING AND CLOUD NATIVE SPACEA. Trends in edge computingB. Infrastructure usage of edge developers4. USAGE TRENDS OF ORCHESTRATION AND SERVERLESS TOOLSMETHODOLOGYTHE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION5

KEY INSIGHTS The global number of cloud native developers has grown in the last 12 months by 0.3 million, to 6.8 million. At the same time, theproportion of backend developers involved in cloud native technologies has declined by 3 percentage points from 44% to 41%. Across regions, North America (47%) and Western Europe (46%) have the highest adoption rates. Overall, larger organisations and moreexperienced developers in these regions are driving adoption. Kubernetes is used by 31% of all backend developers, which is an increase of 4 percentage points in the last 12 months. Currently, 5.6M*developers use Kubernetes. Across sectors, edge computing has experienced rapid growth in the adoption of Kubernetes and now has the highest usage rates of bothcontainers and Kubernetes. Among edge developers only, Kubernetes usage increased in the last 12 months by 11 percentage points to 63%. Edge technology has thehighest adoption rate among all surveyed sectors. Serverless architecture is also appealing to edge developers: 48% of all edge developers are using serverless compared to just 33% of allbackend developers. Among serverless tools, AWS Lambda continues to be in the driving seat. However, Google Cloud Run has been gaining considerably inthe last 12 months.*Our methodology has been updated since our Q1 2020 report to ask all backend developers if they use Kubernetes, rather than just developers who identify as usingcontainer orchestration tools & management platforms. This results in an increase in the number of developers who say they use Kubernetes. In Q1 2020, using thisapproach, we observed 3.9M Kubernetes developers in the total population of backend developers, compared to the stated 2.7M who identified as using containerorchestration tools & management platforms.

1. IntroductionTHE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION

1. INTRODUCTIONA. BackgroundThe way software is developed has drastically changed sincecontainers came about and cloud native technology gainedpopularity. Commissioned by Cloud Native ComputingFoundation (CNCF), SlashData has performed an in-depthanalysis of the cloud native developer ecosystem to betterunderstand its current state and expected evolution. Theanalysis is based on the 20th edition of SlashData’s DeveloperEconomics survey, which ran between November 2020 andFebruary 2021 and reached more than 19,000 developersglobally. 3,800 survey participants answered questionsrelating to the development of backend services and thetechnologies they use.THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION8

1. INTRODUCTIONB. RoadmapAs in our previous report1, we use CNCF’s definition of cloudnative computing as a guide for this analysis:“Cloud native technologies empower organisations to build andrun scalable applications in modern, dynamic environments suchas public, private, and hybrid clouds. Containers, service meshes,microservices, immutable infrastructure, and declarative APIsexemplify this approach. These techniques enable loosely coupledsystems that are resilient, manageable, and observable. Combinedwith robust automation, they allow engineers to make highimpact changes frequently and predictably with minimal toil.”Particular attention is given to edge computing in the thirdchapter of this report. Edge computing naturally intersects withthe cloud native space. We investigate the trends andpreferences of edge developers and the infrastructure that isneeded to facilitate the adoption of cloud native computing.Many use cases of edge computing require applications to resideon-premises or in the cloud, while others run dynamically at theedge. Edge computing thus recommends itself as a paradigm foradopting Kubernetes, which can handle complex systems withease through real-time automation and fast scalability ofcontainers.The first chapter of this report provides estimates for the globalnumber of cloud native developers and an in-depth regionaloverview of usage rates.The report concludes with an in-depth overview of variousorchestration and serverless solutions of selected vendors andhighlights notable trends in this exciting space.Kubernetes is arguably at the heart of cloud native applicationsand in the second chapter, we examine developers' usage andawareness of Kubernetes and attempt to identify the drivers thatfacilitate its adoption across different developer profiles andtechnologies.1The state of cloud native development Q1 2020THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION9

1. INTRODUCTIONC. Market sizeOur estimates put the global number of cloud native developersfor Q1 2021 at 6.8 million, or 41% of backend developers. Thisincludes 4.6M developers using container orchestration tools and4M developers using serverless platforms; the numberscorrespond to 28% and 24% of backend developers,respectively. The estimate also considers an overlap of 1.8Mdevelopers using both orchestration and serverless technologies,corresponding to 11% of backend developers.Although we find that the cloud native developer populationhas grown by 0.3M in the last 12 months, the proportion ofbackend developers who are involved in cloud nativetechnologies has actually decreased slightly by 3 percentagepoints, due to a drop in the adoption of serverless architecture.The growth of the cloud native developer population has beendriven by a robust increase in the total number of backenddevelopers from 14.7M to 16.6M.We previously noted in our Q1 2020 report that developers arereluctant in their adoption of serverless architecture. Recentdata continues to substantiate this trend. While the total numberof developers involved in serverless architecture stagnates ataround 4M, the proportion of developers involved in serverlessarchitecture has decreased from 27% to 24%.Notwithstanding the hype around serverless architecture, whatcauses the lag in its usage? This downward trend appears toreflect the voice of many pundits that claim that serverlesssolutions may lack the flexibility to facilitate widespreadadoption. For instance, companies are reluctant to lockthemselves into specific vendors.THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION10

1. INTRODUCTIONThe cloud native developer population has grown to 6.8 millionActive cloud native developers (in millions)Q1 2020Q3 2020Q1 202110.2 M9.0 M8.2 M6.5 M6.8 M6.1 M4.3 MCloud nativeContainers4.6 M4.1 MContainer orchestration tools &management platforms3.9 M3.9 M4.0MCloud functions or serverlessarchitectureTHE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION11

1. INTRODUCTIOND. Regional usage of cloud native technologiesThe usage of containers and cloud native technologies variessignificantly by region. Despite a significant increase in theadoption rates of containers in the last 12 months in almost allregions - except for the Middle East and Africa - the adoption ofserverless solutions has been slightly dwindling, whereas theadoption of container orchestration remained somewhatunchanged in many regions, including Europe and NorthAmerica.Unlike the usage of containers and container orchestration,serverless solutions have experienced a significant drop in theirusage in many regions, such as North America (-6 percentagepoints), Eastern Europe (-8 percentage points), South Asia (-6percentage points), and the Middle East and Africa (-4percentage points). The two regions that currently defy thenegative trend are Oceania ( 9 percentage points) and SouthAmerica ( 5 percentage points).Notwithstanding, the increase in the adoption of containerscontinues the positive trend that we noticed in our previousreport. Presently, 61% of backend developers use containers.Western Europe boasts the highest usage rates, with three outof four developers having adopted containers, while the lowestadoption rates are found in the Middle East and Africa. Less thanhalf of developers are managing their application processesthrough containers in this region.Overall, cloud native computing is now most widely adopted inOceania (56%), North America (47%), and Western Europe(46%). It remains the least adopted in the Middle East and Africa(28%) and Eastern Europe (37%).There are good reasons for this divergence in usage rates. As wewill see later, the adoption of container orchestration is closelytied to the company size; that is, we observe substantiallyhigher adoption rates within larger organisations, likely driven byproject complexity. When excluding small organisations with 10or fewer employees from our analysis, cloud native adoptionrates become more balanced across regions, for example,Western Europe (49%), the Middle East and Africa (45%), andEastern Europe (43%).THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION12

1. INTRODUCTIONUsage of cloud native technologies has been shrinking in NorthAmerica, Eastern Europe, and the Middle East and Africa% of backend developers in each region (Q1 2020 n 4,179 Q1 2021 n 3,836)Usage of cloud native technologies across regionsQ1 202173%62%Q1 stern Europe& Israel32%32%11%19%37%17%43%Eastern Europe,Russia & FormerCIS28%21%34%Middle East &Africa23%37%22%39%East Asia22%26%38%23%32%43%South Asia30%47%36%50%North America28%32%29%25%38%27%29%22%56%41%42%South America41%OceaniaTHE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION13

2. USAGE ANDAWARENESS OFKUBERNETESTHE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION

2. USAGE AND AWARENESS OF KUBERNETESA. Kubernetes and containers usage and awareness amongbackend developersWe asked around 7,000 backend developers about theirawareness and usage of both Kubernetes and containers. Thislarge sample comprised both cloud native and non-cloud nativedevelopers and provides an unbiased picture of how well-knownthese cloud native technologies are among the whole populationof backend developers.This section explores backend developers’ awareness and usageof Kubernetes and containers and examines interesting trends inthe data. As of Q1 2021, 57% of backend developers had usedcontainers in the last 12 months, but only 31% of developers usedKubernetes to orchestrate these: 5.6M developers in total. Theiroverall usage of Kubernetes has increased by 4 percentagepoints in the last 12 months. Kubernetes thus seems to exhibit adistinctive positive trend within the cloud native space, andthere is arguably still room to grow.However, 21% of backend developers say they have heard ofKubernetes but are unsure what it does. Another 11% say theyhave not heard of Kubernetes. While overall awareness ofKubernetes has significantly increased, many backenddevelopers remain unsure of what it can do for them. To shedmore light on this matter, we will next examine in more depththe profiles of developers who have adopted Kubernetes.It’s important to note that we updated our methodology the Q12020 report to ask all backend developers if they useKubernetes, rather than just developers who identify as usingcontainer orchestration tools & management platforms. Thischange resulted in an increase in the number of developers whosay they use Kubernetes. In Q1 2020, using this approach, weobserved 3.9M Kubernetes developers in the total population ofbackend developers, compared to the stated 2.7M who identifiedas using container orchestration tools & management platforms.THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION15

2. USAGE AND AWARENESS OF KUBERNETESKubernetes and containers usage and awareness among backenddevelopers% of backend developers who have used each technology in the past 12 months (Q1 2020 n 6,554 Q1 2021 n 6,927)Heard of it but not sure what itdoesUsed in the past 12 monthsKubernetes58%27%Never heard of itContainers57%31%23%21%13%14%15%11%5%Q1 2020Q1 2021Q1 2020Q1 2021Q1 20206%Q1 2021THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION16

2. USAGE AND AWARENESS OF KUBERNETESA. Kubernetes and containers usage and awareness amongbackend developersWe further examined usage of Kubernetes and containers bysixteen technologies that developers are currently working on.For example, we considered technologies such as fog and edgecomputing, 5G, robotics, quantum computing, cryptocurrencies,and other blockchain applications, among others.Developers working on edge computing have the highestusage for both containers and Kubernetes. Containers havebeen used here in the last 12 months by 76% and Kubernetes by63% of developers currently working on edge computingapplications.On the flip side, the technologies with the smallest headroom arehaptic feedback1 (5 percentage points), quantum computing (6percentage points), and self-driving cars (8 percentage points).Why are developers in certain technologies more reluctant toadopt Kubernetes despite the high usage of containers? At firstglance, the technologies providing less headroom seem moredemanding or complex in terms of development needs, forexample, as presumed in the cases of quantum computing orhaptic feedback.While there is arguably some headroom to adopt Kubernetes inedge computing - there’s a 13 percentage point gap betweencontainers and Kubernetes in terms of adoption - there are othertechnologies worth mentioning here. The largest headrooms arein technologies that involve the development of mini apps andcomputer vision, with 27 percentage point and 23 percentagepoint gaps between the usage of containers and Kubernetes,respectively.1Hapticfeedback is the use of touch feedback to end-users through vibrations or motions.THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION17

2. USAGE AND AWARENESS OF KUBERNETESKubernetes and container usage and awareness among backenddevelopers% of backend developers using containers and Kubernetes by area of involvement (Q1 2021 n 6,927)Usage of Kubernetes and containers in the last 12 months by technologyFog / edge computingQuantum computingHaptic feedbackDNA computing / storage5GBlockchain applications other than cryptocurrencyHearablesBrain / body computer interfacesSelf-driving carsDronesComputer visionBiometrics for ID verificationConversational platforms / voice searchCryptocurrencies (e.g. Bitcoin)RoboticsMini ContainersKubernetesTHE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION18

2. USAGE AND AWARENESS OF KUBERNETESB. Profile of backend developers using KubernetesTo test the claim of the complexity of the technology drivingKubernetes usage, we segmented the data by developers’professional status, experience, and the size of the organisationthey are working in and ran a statistical model. The mostsignificant predictor that drives Kubernetes adoption is thesize of the organisation that a developer works for and thisrelationship holds regardless of the fact that larger companieshave, in general, more experienced developers working forthem.Indeed, the adoption of both Kubernetes and containersskyrockets for large organisations with more than 500employees. Moreover, the average headroom across alltechnologies decreases from 20% for developers in small andmedium-sized organisations to just 11% for developers working inlarge organisations. In some technologies, Kubernetes adoptioneven matches container usage. For example, 83% of developersworking on haptic feedback technology, i.e., the use of touchfeedback to end-users through vibrations or motions, in largeorganisations use both containers and Kubernetes.Interestingly, the difference between container and Kubernetesusage is just 5% within large organisations with high adoptionrates for both containers (83%) and Kubernetes (78%).Therefore, within larger organisations, the adoption ofKubernetes among edge developers is already high relative tothe adoption of containers. It stands to reason that edgedevelopers are also more likely to be employed in largerorganisations, for example, in the industrial IoT sector.The technologies that have the most considerable differencebetween container and Kubernetes usage among developersworking in large companies are self-driving cars (19 percentagepoints), biometrics for ID verification (19 percentage points),blockchain applications other than cryptocurrency (18percentage points), and 5G (17 percentage points).The adoption of Kubernetes arguably requires a significantinvestment of time and resources with a cluster of severalmachines by default. The data backs up a strong positivecorrelation between the Kubernetes adoption relative to theadoption of containers and technologies that are likely to betargeted in larger organisations. Large organisations involved incomplex technologies have better means and use cases to adoptKubernetes.THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION19

2. USAGE AND AWARENESS OF KUBERNETESContainer usage is greater in large companies and and is reachinga plateau% of backend developers using Kubernetes by area of involvement (Q1 2021 n 4,442)Company size and containers usage83%86%88%77%77%77%Fog / edge computingBlockchain applications other than cryptocurrencyDNA computing / ini appsQuantum computingComputer visionCryptocurrencies (e.g. Bitcoin)Haptic feedbackRobotics51%63%Conversational platforms / voice searchHearables59%5GSelf-driving cars47%Companies with up to 500 employees69%84%54%52%Large companies with more than 500 employees81%59%Biometrics for ID verificationBrain / body computer interfaces76%61%77%83%THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION20

2. USAGE AND AWARENESS OF KUBERNETESKubernetes usage skyrockets in large companies, but there are stillconsiderable variations across areas% of backend developers using Kubernetes by area of involvement (Q1 2021 n 4,442)Company size and Kubernetes usageHaptic feedback83%47%Fog / edge computingQuantum computing52%Blockchain applications other than cryptocurrency70%39%5G67%45%Drones66%42%Brain / body computer interfaces43%Self-driving cars43%65%64%64%59%63%DNA computing / storageCryptocurrencies (e.g. Bitcoin)Conversational platforms / voice searchRoboticsMini appsBiometrics for ID blesComputer vision67%71%Large companies with more than 500 employees53%53%Companies with up to 500 employees50%THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION21

3. EDGECOMPUTING ANDCLOUD NATIVESPACETHE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION

3. EDGE COMPUTING AND CLOUD NATIVE SPACEA. Trends in edge computingIn the previous chapter, we noted that developers who work onedge computing are more likely to rely on both containers andKubernetes. In this chapter, we go a level deeper and look at thedetailed usage profile of edge developers and bring out trendsat the intersection between edge and cloud native spaces.As adoption of containers has not been trending upwards, it maybe a sign that it is already or slowly reaching a high plateau; thatis, a point where further adoption becomes less likely. On theother hand, the usage of Kubernetes appears to still have plentyof room to grow.At first glance, we see that container usage, for both backenddevelopers in general and those working on edge computingprojects, has been rather unchanged in the last 12 months. At thesame time, the adoption of Kubernetes among backenddevelopers grew by 4 percentage points. Moreover, amongedge developers only, the adoption of Kubernetes increasedby 11 percentage points.THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION23

3. EDGE COMPUTING AND CLOUD NATIVE SPACETwo in three edge developers now use Kubernetes% of backend developers who have been involved in each area in the past 12 months(Q1 2020 n 131 Q3 2020 n 151 Q1 2021 n 106)Containers usageAll backend developersKubernetes usageFog and edge developers77%56%Q3 202063%53%57%26%Q1 2020Fog and edge developers60%76%71%58%All backend developersQ1 2021Q1 202027%Q3 202031%Q1 2021THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION24

3. EDGE COMPUTING AND CLOUD NATIVE SPACEA. Trends in edge computingThe adoption of Kubernetes increases in larger organisations,and, in the space of edge computing, almost matches theadoption of containers. Yet, growth in adoption can be stillachieved twofold:Increasing awareness of and exploring use cases for Kubernetesfor small and medium-sized organisations. Although theadoption of Kubernetes among edge developers in small andmedium-sized organisations has been continuously growing andcurrently stands at 60%, the data shows that 17% of edgedevelopers in small companies are still not sure whatKubernetes does - a figure that has doubled within a year.When it comes to the use of serverless technologies, 48% ofedge developers currently go serverless compared to only 33%of all backend developers. The lightweight nature of a serverlessarchitecture is particularly appealing to edge developers sincethey don’t have to manage the underlying infrastructure.However, serverless usage rates have grown at a slower pacethan Kubernetes and recently even experienced an unexpecteddrop. A drop in serverless usage has also been noticeable amongall backend developers and, while the cause of it is not fullyunderstood, it remains to be seen if it will constitute a morepermanent shift.Even if the Kubernetes adoption rate reaches a plateau bymatching the container adoption rate, the underlying growth inthe sector will continue to propel the total usage of Kubernetes.For example, IoT, arguably the key driver of many edgesolutions, has been on a steep trajectory. The IoT sector isexpected to grow by approximately 19% per year, as forecast inour IoT population report.THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION25

3. EDGE COMPUTING AND CLOUD NATIVE SPACEServerless solutions experienced a recent drop in usage% of backend and edge developers(Q1 2019 n 2,288 Q3 2019 n 1,868 Q1 2020 n 1,817 Q3 2020 n 1,752 Q1 2021 n 870)Usage of serverless architectureEdge developers56%54%54%53%All backend developers48%30%Q1 201934%Q3 201936%Q1 202036%Q3 202033%Q1 2021THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION26

3. EDGE COMPUTING AND CLOUD NATIVE SPACEB. Infrastructure usage of edge developersIn our survey, we asked backend developers where their coderuns, with potential answers including public cloud, privatecloud, hybrid cloud, multi-cloud, and on-premise servers. Theseoptions are not mutually exclusive. For greater clarity, within thesurvey, we define a private cloud as a cloud that is only availableto certain users regardless of whether it is hosted on-premise orin a third party's data centre. We also define hybrid clouds asusing a combination of public and private clouds for a singleproject and multi-clouds as using multiple public clouds for asingle project.While the average backend developers who use Kubernetesseem less concerned with the choice of the infrastructure thathosts their code, an edge developer shows a clear preferenceof private clouds (64%), public clouds (63%), and on-premisesservers (59%) over multi-clouds (44%) and hybrid clouds (42%).The figures suggest that developers working on edgeapplications relying on Kubernetes gravitate towards a moresecure and less complex infrastructure. Vast volumes of sensitivedata are handled over much larger distances and achievingreliable and secure connectivity between providers comes withmajor challenges, while varying functionality and user experienceof each cloud platform pose roadblocks on their own.THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION27

3. EDGE COMPUTING AND CLOUD NATIVE SPACEHybrid and multi-cloud are less popular among Kubernetes users% of backend developers (n 1,874)Where do Kubernetes users run their code?Backend developers64%63%Edge developers59%49%40%Private cloud (cloud onlyavailable to certain users)42%44%46%42%35%Public cloudOn-premise serversMulti-cloud (using multiplepublic clouds for a singleproject)Hybrid cloud (using publicand private clouds for asingle project)THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION28

3. EDGE COMPUTING AND CLOUD NATIVE SPACEB. Infrastructure usage of edge developersNext, we look at the cloud vendors that Kubernetes users prefer.Here, we compare edge developers’ vendor preferences to thoseof traditional backend developers.Amazon Web Services (AWS) is the most widely used cloudservice provider among backend developers who rely onKubernetes, with a dominant share of 64%. Their lead is trailedby Google Cloud Platform (45%) and Microsoft Azure (42%). Sofar, the results only reflect the findings in our previous report;however, an interesting pattern surfaces when we furthersegment the data by edge developers.While the three top vendors show a similar usage for developersworking on edge applications, self-hosted solutions and manyother players are significantly gaining popularity. For example,42% of edge developers using Kubernetes prefer a self-hostedsolution compared to only 33% of traditional developers.Similarly, IBM Cloud becomes much more appealing among edgedevelopers (35%) than when compared to the average backenddeveloper (15%). The vendor landscape becomes overall morebalanced within the space between cloud native computingand edge.THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION29

3. EDGE COMPUTING AND CLOUD NATIVE SPACEKubernetes together with edge flattens the usage landscape ofvendors% of backend developers using Kubernetes (n 1,244)Usage of cloud vendors61%Amazon Web Services (AWS)Microsoft Azure42%Google Cloud Netlify31%24%OracleVultr32%14%HerokuSAP35%15%Red HatOVH

THE STATE OF CLOUD NATIVE DEVELOPMENT Q1 2021 KEY INSIGHTS FOR THE CLOUD NATIVE COMPUTING FOUNDATION The cloud native developer population has grown to 6.8 million 6.5 M 8.2 M 4.3 M 3.9 M 6.1 M 9.0 M 4.1 M 3.9 M 6.8 M 10.2 M 4.6 M 4.0M Cloud native Containers Container orchestration tools & m