Java Tech Popularity Index Q4/2023: Back-End Frameworks

Summary for Q4/2023

Here is the scorecard of Spring Boot (left) and Quarkus (right) vs. Jakarta EE (100%), not on the card. The arrows show the trend vs. Jakarta EE.

Spring Boot pulls away from Jakarta EE everywhere except for job ad mentions. Quarkus gains on Jakarta EE everywhere.

Here are my recommendations:

• On your current project, keep your existing back-end framework unless that framework is absolutely, really not working out for you.
• If you need to switch back-end frameworks or are on a new project:
  • Use Quarkus if you need the smallest possible, fastest-starting Java application now.
  • Otherwise, use Spring Boot.

Archive

Table Of Contents

• Summary for Q4/2023
• Archive
• Choices
• Popularity
  • Why Popularity - and How?
  • Employers: Job Ads
  • Developers
    • Courses Bought at Udemy
    • Google Searches
    • Monthly Questions at Stack Overflow
• Analysis
• Next Issue

Choices

Here are the choices in alphabetical order:

• Dropwizard
• Oracle’s Helidon
• Eclipse Jakarta EE (was Java EE before)
• Micronaut
• Eclipse MicroProfile
• VMWare’s Spring Boot
• Red Hat’s Quarkus

Popularity

Why Popularity - and How?

Picking a popular technology makes our developer life easier: Easier to learn, easier to build, debug & deploy, easier to find jobs/hire, and easier to convince teammates & bosses. Popularity can make a difference in two situations: When multiple technologies score similarly, we could go for the most popular one. And when a technology is very unpopular, we may not use it.

I measure popularity among employers and developers as the trend between competing technologies. I count mentions in job ads at Indeed for employer popularity. For developer popularity, I use Google searches, Udemy course buyers, and Stack Overflow questions.

Employers: Job Ads

The Indeed job search is active in 62 countries. I picked 59 countries representing 69% of the worldwide GDP in 2022, excluding three countries because English word searches proved ineffective there: China, Japan, and South Korea. Job searches demonstrate the willingness of organizations to pay for a technology - the strongest indicator of popularity in my mind. Jakarta EE is the baseline. The chart is not proportional, so all frameworks fit nicely.

There are no job ad numbers for May 2023 because of changes on the Indeed websites.

Spring Boot wins, Jakarta EE is second, Quarkus third, and Micronaut fourth. The overall number of developer ads is down in 2023. Spring Boot has lost 13% in the last year, dropping to a 5.1:1 lead over Jakarta EE. The new kids, Quarkus and Micronaut, at #3 and #4, saw their rebound continue. After a rollercoaster ride, Dropwizard went from #3 to #5 at 36% of its starting share. Helidon and MicroProfile are insignificant in the job market.

Please see here for details, caveats, and adjustments to the job ad mentions.

You can find the detailed search results with links here. They include breakdowns by continents:

Developers

Courses Bought at Udemy

Udemy is one of the biggest online learning sites. They publish the number of people who bought a course (beyond a certain threshold, possibly around 100k). This shows how many people evaluate a technology. Jakarta EE is the baseline. The other frameworks haven’t crossed the reporting threshold for Udemy (probably around 100,000 students).

Spring Boot wins over Jakarta EE and increases its lead from 3.6:1 to 5.1:1. The Spring Framework also grows strongly but loses to Spring Boot. In April this year, Udemy actually decreased the number of Jakarta EE students from 242k to 230k – that’s the small jump in share for Spring and Spring Boot.

Here are the links that show the courses for all and the number of students for some:

• Dropwizard
• Helidon
• Jakarta EE
• Micronaut
• MicroProfile
• Spring Boot
• Quarkus

Google Searches

Google Trends demonstrates the initial interest in a technology over time. “More searches = better” to me. The percentage behind the current value is the drop-off from the peak value, marked with a circle. Here are all the frameworks, but Helidon - Google Trends only allows five at the same time:

Google changed its measurement algorithms on January 1, 2016, and January 1, 2022. That caused spikes for all values, especially in 2022.

This link produces the chart above. This version switches in Helidon for Micronaut and this one MicroProfile — which isn’t making a difference in the chart.

Spring Boot wins, Jakarta EE is second, Quarkus third, Micronaut fourth, and DropWizard fifth. Jakarta EE lost 98% since 2004, equaling Quarkus at two today. Spring Boot is still declining and is 15% off its all-time high. It leads Quarkus and Jakarta EE 17:1. Quarkus just hit its peak value, as did #4 Micronaut. DropWizard is below 1.

Let’s zoom in on the five challengers over the last three years (chart link):

Quarkus leads the new frameworks. Quarkus has more than doubled in the last three years and may grow again after sliding earlier this year. Micronaut and Helidon grew 60% over the previous three years. As the #2 challenger, Micronaut has 29% of Quarkus’ search volume.

How does Jakarta EE fare against Quarkus over the last three years (chart link)?

Jakarta EE beats Quarkus narrowly. Jakarta EE currently leads Quarkus by a hair. The trend suggests that Quarkus may soon overtake Jakarta EE.

Monthly Questions at Stack Overflow

We can run database queries against the questions, answers, and comments at Stack Overflow with the StackExchange Data Explorer. The number of monthly questions is a proxy for using a technology during evaluation and productive use. “More questions = better” to me. The percentage behind the current value is the drop-off from the peak value, marked with a circle.

Spring Boot wins by an order of magnitude, while Quarkus places second and Jakarta EE third. Spring Boot lost 50% of its question volume in the last three years, and 40% since March. Quarkus weathered the general Stack Overflow decline much better with 102 questions, being off just 11% of its peak value. Micronaut’s 20 questions (per month!) are enough for the #3 spot. Jakarta EE is down to a mere 11 questions, 2% of its all-time high of 486 questions in 2013. DropWizard shrunk from 51 questions in 2015 to just five last month. Helidon managed a single question in September.

Please note that the overall monthly number of Stack Overflow questions is down 42% since ChatGPT appeared (November 2022 vs. October 2023):

I used two queries to get the number of monthly questions below because putting all in one query timed out. You can run them in the StackExchange Data Explorer.

DECLARE @StartDate DATE = '2009-01-01';
DECLARE @EndDate DATE = '2023-09-30';

WITH TaggedQuestions AS (
  SELECT
    Id,
    CreationDate,
    CASE WHEN CHARINDEX('<spring>', Tags) > 0 THEN 1 ELSE 0 END AS SpringTag,
    CASE WHEN CHARINDEX('<spring-boot>', Tags) > 0 THEN 1 ELSE 0 END AS SpringBootTag,
    CASE WHEN CHARINDEX('<jakarta-ee>', Tags) > 0 THEN 1 ELSE 0 END AS JakartaEETag
  FROM Posts
  WHERE
    PostTypeId = 1 AND -- 1 for questions
    CreationDate >= @StartDate AND
    CreationDate <= @EndDate
),
MonthlyCounts AS (
  SELECT
    DATEADD(month, DATEDIFF(month, 0, CreationDate), 0) AS Month,
    SUM(SpringTag) AS Spring,
    SUM(SpringBootTag) AS SpringBoot,
    SUM(JakartaEETag) AS JakartaEE
  FROM TaggedQuestions
  GROUP BY DATEADD(month, DATEDIFF(month, 0, CreationDate), 0)
)
SELECT
  Month,
  Spring,
  SpringBoot,
  JakartaEE
FROM MonthlyCounts
ORDER BY Month;

DECLARE @StartDate DATE = '2009-01-01';
DECLARE @EndDate DATE = '2023-09-30';

WITH TaggedQuestions AS (
  SELECT
    Id,
    CreationDate,
    CASE WHEN CHARINDEX('<quarkus>', Tags) > 0 THEN 1 ELSE 0 END AS QuarkusTag,
    CASE WHEN CHARINDEX('<micronaut>', Tags) > 0 THEN 1 ELSE 0 END AS MicronautTag,
    CASE WHEN CHARINDEX('<helidon>', Tags) > 0 THEN 1 ELSE 0 END AS HelidonTag,
    CASE WHEN CHARINDEX('<dropwizard>', Tags) > 0 THEN 1 ELSE 0 END AS DropWizardTag
  FROM Posts
  WHERE
    PostTypeId = 1 AND -- 1 for questions
    CreationDate >= @StartDate AND
    CreationDate <= @EndDate
),
MonthlyCounts AS (
  SELECT
    DATEADD(month, DATEDIFF(month, 0, CreationDate), 0) AS Month,
    SUM(QuarkusTag) AS Quarkus,
    SUM(MicronautTag) AS Micronaut,
    SUM(HelidonTag) AS Helidon,
    SUM(DropWizardTag) AS DropWizard
  FROM TaggedQuestions
  GROUP BY DATEADD(month, DATEDIFF(month, 0, CreationDate), 0)
)
SELECT
  Month,
  Quarkus,
  Micronaut,
  Helidon,
  DropWizard
FROM MonthlyCounts
ORDER BY Month;

DECLARE @StartDate DATE = '2009-01-01';
DECLARE @EndDate DATE = '2023-10-31';

WITH Questions AS (
  SELECT
    Id,
    CreationDate,
    1 AS AQuestion
  FROM Posts
  WHERE
    PostTypeId = 1 AND -- 1 for questions
    CreationDate >= @StartDate AND
    CreationDate <= @EndDate
),
MonthlyCounts AS (
  SELECT
    DATEADD(month, DATEDIFF(month, 0, CreationDate), 0) AS Month,
    SUM(AQuestion) AS Questions
  FROM Questions
  GROUP BY DATEADD(month, DATEDIFF(month, 0, CreationDate), 0)
)
SELECT
  Month,
  Questions
FROM MonthlyCounts
ORDER BY Month;

Analysis

• Spring Boot dominates the Java ecosystem. It has the broadest support of libraries and third-party software working out of the box, conveniently configured consistently with Spring Boot. Spring Boot 3 now has built-in support to create native Java executables with the GraalVM Native Image AOT compiler. You know, just like Quarkus, Micronaut, and Helidon have done for years. But with a multi-year head-start, these Spring Boot competitors probably produce smaller and faster native Java executables. If that matters to you, then go Quarkus or Micronaut.
• Jakarta EE is a vendor-independent specification with multiple implementations. It has had a rough couple of years: First, Oracle neglected it when it was still called “Java EE”. Oracle grudgingly transferred it to the Eclipse Foundation but required renaming all packages from javax.* to jakarta.*. After Java EE 8 in August 2017, we had to wait five years for new features in Jakarta EE 10. But the more significant issue is that Jakarta EE was designed for application servers like IBM WebSphere that host many applications on big and expensive servers. We’re rapidly moving into a world where our Java applications run as close to the metal as possible, as microservices in small containers or even serverless. Will that world still have a place for Jakarta EE? Only time will tell.
• MicroProfile is “Spring Boot with more Jakarta EE parts”, but without its broad support of libraries and third-party software. It started out of frustration with the slow progress of Java EE in 2016. Just like Jakarta EE, it is a vendor-independent specification with multiple implementations. It’s probably the least popular framework here, based on my measurements.
• Dropwizard is a framework that I had never heard of until the “JRebel 2021 Java Technology Report” declared it the #2 Java framework. Its origins seem to go back to 2012. It has the most job ad mentions besides Spring Boot and Jakarta EE.
• Quarkus is Red Hat’s take on a cloud-native Java framework. “Cloud-native” means producing small and fast applications: Quarkus claims 12 MB RAM for a REST application that starts in 0.016 seconds. It has the most mind-share as the cloud-native Spring Boot competitor. It’s the most popular framework besides Spring Boot and Jakarta EE.
• Micronaut is the second cloud-native Spring Boot competitor. It’s backed by consulting company Object Computing, which sponsors Grails and is less popular than Quarkus.
• Helidon is the third and least popular cloud-native Spring Boot competitor. It’s an Oracle framework that is rather unpopular.

So here’s my recommendation:

• On your current project, keep your existing back-end framework unless that framework is absolutely, really not working out for you.
• If you need to switch back-end frameworks or are on a new project:
  • Use Quarkus if you need the smallest possible, fastest-starting Java application now.
  • Otherwise, use Spring Boot 3.

Next Issue

The next issue will arrive in November 2023. Subscribe to it as a newsletter to have it in your inbox then!