This Java 21 project benchmarks a simple Spring Boot 3.2 HTTP endpoint using configurable scenarios, comparing Java Virtual Threads (introduced by Project Loom, JEP 444) using Tomcat and Netty with Spring WebFlux (relying on Project Reactor) using Netty.
All benchmark results below come from a dedicated bare metal test environment. The benchmark is also scheduled to run monthly on a GitHub-hosted runner, using both Ubuntu 22.04 and 24.04.
Both Spring WebFlux and Virtual Threads are alternative technologies to create Java microservices that support a high number of concurrent users, mapping all incoming requests to very few shared operating system threads. This reduces the resource overhead incurred by dedicating a single operating system thread to each user.
Spring WebFlux was first introduced in September 2017. Virtual Threads were first introduced as preview feature with Java 19 and were fully rolled out with Java 21 in September 2023.
Note
In a nutshell, the benchmark results are:
Virtual Threads on Netty (using blocking code) showed very similar and often superior performance characteristics (latency percentiles, requests per second, system load) compared with WebFlux on Netty (using non-blocking code and relying on Mono and Flux from Project Reactor):
- Virtual Threads on Netty was the benchmark winner for about 75% of all combinations of metrics and benchmark scenarios when directly compared with Project Reactor on Netty.
- For all high user count scenarios, it had the lowest latency as well as the largest number of requests for the entirety of each benchmark run.
- In many cases (e.g. 60k-vus-smooth-spike-get-post-movies), the 90th and 99th percentile latencies (P90 and P99) were considerably lower for Virtual Threads on Netty when compared with WebFlux on Netty.
- For both approaches, we could scale up to the same number of virtual users (and thus TCP connections) before exhausting the CPU and running into time-outs due to rejected TCP connection requests.
Virtual Threads on Tomcat are not recommended for high load:
- We saw considerably higher resource use compared with the two Netty-based approaches.
- There were many time-out errors as visualized by red dots in the charts, even when the CPU use was far below 100%. In contrast, none the Netty-based scenarios experienced any errors, even with a CPU use of 100%.
Below are top-performing approaches across all scenarios and metrics, visualizing the contents of results/results.csv:
- Each cell shows the metric values of best approach (on top) and runner-up.
- What "best" is depends on the metric: A lower value is better for all metrics except for metrics starting with
requests_ok,requests_per_second, orsockets. - Approaches which encountered request errors are ranked below those approaches with only successful requests. Such "failed" approaches have all their metric values printed in red and suffixed with
E. - An overall ranking based on the win count of each approach is shown in the legend:
(1)indicates the overall best approach,(2)the runner-up, and so on. This overall ranking is also shown next to each metric value.
- What "best" is depends on the metric: A lower value is better for all metrics except for metrics starting with
- Cells are colored based on the winning approach. The darker the color, the bigger the lead of the winning approach over the runner-up. If cells are white or faded, there's no clear winner as the top two approaches performed similarly.
- For detailed charts on each approach and test scenario combination, have a look at the second half of this document.
This chart compares Project Loom (on both Tomcat and Netty) with Project Reactor (on Netty).
This chart is based on same benchmark as before, but only considers Netty-based approaches.
- Fully automated and CLI-driven by running a single command:
benchmark.sh. - Different test scenario files, each containing one or more scenarios. Example:
src/main/resources/scenarios/scenario.csv. - Operating system thread re-use by waiting and by performing transitive HTTP calls of configurable call depth.
- Interacts with realistic JSON APIs.
- Creates single PNG image via Matplotlib for each combination of scenario and approach which contains:
- Raw latencies and P50/90/99 percentiles, as well as any errors.
- System metrics for CPU, RAM, sockets, and network throughput.
- JVM metrics such as heap usage, garbage collections (GCs), and platform thread count.
- Creates summary PNG image of all scenarios which shows best approaches.
The benchmark is driven by k6 which repeatedly issues HTTP requests to a service listening at http://localhost:8080/
The service exposes multiple REST endpoints. The implementation of each has the same 3 stages:
- HTTP Call: If
$delayCallDepth > 0, callGET /$approach/epoch-millisrecursively$delayCallDepthtimes to mimic calls to upstream service(s).- All approaches use
Spring Boot's WebFlux WebClient based on Netty.
- All approaches use
- Wait: If
$delayCallDepth = 0, wait$delayInMillis(default:100) to mimic the delay incurred by a network call, filesystem access, or similar.- Whilst the request waits, its operating system thread can be reused by another request.
- The imperative approaches (
platform-tomcat,loom-tomcat, andloom-netty) use blocking wait whilst the reactive approach (webflux-netty) uses non-blocking wait.
- Calculate and Return Response specific to REST endpoint.
Get all movies using loom-netty approach, an HTTP call depth of 1 and a delay of 100 milliseconds:
sequenceDiagram
participant k6s
participant service
k6s->>+service: GET /loom-netty/movies?delayCallDepth=1&delayMillis=100
service->>+service: GET /loom-netty/epoch-millis?delayCallDepth=0&delayMillis=100
service->>service: Wait 100 milliseconds
service-->>-service: Return current epoch millis
service->>service: Find movies
service-->>-k6s: Return movies
The microservice under test exposes several RESTful APIs. In the following descriptions, $approach is the approach
under test and can be one of loom-tomcat, loom-netty, and webflux-netty.
All REST APIs support the following query parameters:
delayCallDepth: Depth of recursive HTTP call stack to$approach/epoch-millisendpoint prior to server-side delay; see Scenario Columns for more details.delayInMillis: Server-side delay in milliseconds; see Scenario Columns for more details.
The TimeController returns the milliseconds since the epoch, i.e. 1 Jan 1970:
- This is one of the simplest possible APIs to provide a best-case performance scenario.
- Supported requests:
GET /$approach/epoch-millis
The MovieController gets and saves movies which are stored in an in-memory repository:
- This is a realistic JSON API as exposed by a typical microservice.
- Several hard-coded movies by three directors are provided.
- By default, writes are not saved since the code under test is identical for all approaches and would thus only
contribute to CPU use. However, this can be controlled with the Spring Boot property
loom-webflux.repo-read-onlyinsrc/main/resources/application.yaml. - Supported requests:
GET /$approach/movies?directorLastName={director}:- Returns movies by the specified director.
- Supported
{director}values and their respective response body size in bytes, based on the default movies:Allen: 1597 bytes (unindented)Hitchcock: 1579 bytes (unindented)Kubrick: 1198 bytes (unindented)
POST /$approach/movies:- Saves one or more movies.
- The sample movies saved during the load tests measure 7288 bytes (indented).
- Unix-based OS; tested with Ubuntu 22.04
- Java 21 or above
- k6 and Python 3 with Matplotlib to drive load and measure latency
- sar/sadf to measure system resource use
- Python 3 and Matplotlib to convert latency and system CSV measurements into a PNG image
The hardware requirements depend purely on the scenarios configured in src/main/resources/scenarios/scenarios.csv. The following is
recommended to run the default scenarios committed to this repo:
- CPU comparable to Intel 6700K or above
- 16 GiB RAM
The following instructions assume you are using a Debian-based Linux such as Ubuntu 22.04.
You'll need Java 21 or above:
sudo apt install openjdk-21-jdkk6 is used to load the service:
sudo gpg -k
sudo gpg --no-default-keyring --keyring /usr/share/keyrings/k6-archive-keyring.gpg --keyserver hkp://keyserver.ubuntu.com:80 --recv-keys C5AD17C747E3415A3642D57D77C6C491D6AC1D69
echo "deb [signed-by=/usr/share/keyrings/k6-archive-keyring.gpg] https://dl.k6.io/deb stable main" | sudo tee /etc/apt/sources.list.d/k6.list
sudo apt-get update
sudo apt-get install k6Python 3 and matplotlib are used to convert the CSV output of k6 and sar/sadf to a single PNG chart. The sar
and sadf tools come as part of sysstat and are used to measure resource use. To install them run:
sudo apt update && sudo apt install -y python3 python3-matplotlib sysstatThe following adjustments optimize Linux for HTTP load tests.
Ensure your system can handle a large number of concurrent connections:
printf '* soft nofile 1048576\n* hard nofile 1048576\n' | sudo tee -a /etc/security/limits.conf Increase the port range for outgoing TCP connections and allow quick connection reuse:
printf 'net.ipv4.ip_local_port_range=1024 65535\nnet.ipv4.tcp_tw_reuse = 1\n' | sudo tee -a /etc/sysctl.conf && sudo sysctl -pLog out and back in.
The following command runs the benchmark for each combination of approaches and scenarios. Results are stored in
the build/results folder:
./benchmark.sh To see the available options, run benchmark.sh -h:
Usage: benchmark.sh [-h] [-a <approaches>] [-C] [FILE]
FILE: Scenario CSV file. Default: src/main/resources/scenarios/scenarios.csv
-a <approaches>: Comma-separated list of approaches to test. Default: loom-tomcat,loom-netty,webflux-netty
Supported approaches: platform-tomcat,loom-tomcat,loom-netty,webflux-netty
-C Keep CSV files used to create chart. Default: false
-h Print this help
- platform-tomcat: Platform threads using Tomcat server
- loom-tomcat: Virtual Threads using Tomcat server
- loom-netty: Virtual Threads on Netty server
- webflux-netty: WebFlux on Netty server
All approaches use the same Spring Boot 3.2 version.
see src/main/resources/scenarios/scenarios.csv
| Scenario | Domain | Description | Virtual Users (VU) | Requests per Second (RPS) | Client delay (ms) | Server delay (ms) | Delay Call Depth |
|---|---|---|---|---|---|---|---|
| smoketest | Time | Smoke test | 5 | 5 | 0 | 100 | 0 |
| 5k-vus-and-rps-get-time | Time | Constant users, constant request rate | 5,000 | 5,000 | 0 | 100 | 0 |
| 5k-vus-and-rps-get-movies | Movies | Constant users, constant request rate | 5,000 | 5,000 | 0 | 100 | 0 |
| 10k-vus-and-rps-get-movies | Movies | Constant users, constant request rate | 10,000 | 10,000 | 0 | 100 | 0 |
| 10k-vus-and-rps-get-movies-call-depth-1 | Movies | Constant users, constant request rate | 10,000 | 10,000 | 0 | 100 | 1 |
| 25k-vus-stepped-spike-get-movies | Movies | Stepped user spike | 0 - 25,000 | Depends on users and delays | 1000 - 3000 (random) | 100 | 0 |
| 25k-vus-smooth-spike-get-movies | Movies | Smooth user spike | 0 - 25,000 | Depends on users and delays | 1000 - 3000 (random) | 100 | 0 |
| 25k-vus-smooth-spike-get-post-movies | Movies | Smooth user spike | 0 - 25,000 | Depends on users and delays | 1000 - 3000 (random) | 100 | 0 |
| 25k-vus-smooth-spike-get-post-movies-call-depth-1 | Movies | Smooth user spike | 0 - 25,000 | Depends on users and delays | 1000 - 3000 (random) | 100 | 1 |
see src/main/resources/scenarios/scenarios-high-load.csv
| Scenario | Domain | Description | Virtual Users (VU) | Requests per Second (RPS) | Client delay (ms) | Server delay (ms) | Delay Call Depth |
|---|---|---|---|---|---|---|---|
| 60k-vus-smooth-spike-get-post-movies | Movies | Smooth user spike | 0 - 60,000 | Depends on users and delays | 1000 - 3000 (random) | 100 | 0 |
The benchmark run for each $scenario consists of the following phases and steps:
- Build and start the Spring Boot service with a specific
$approachas Spring Boot profile, using the config insrc/main/resources/application.yamland overridden bysrc/main/resources/application-$approach.yamlif defined.
- Run the benchmark as configured by the
$scenario. - For each
$resultType(i.e.latency,system, orjvm), create a CSV file atbuild/results/$scenario/$approach-$resultType.csv.
- Convert CSV files into
build/results/$scenario/$approach.png - Delete the CSV files unless the
-CCLI option was specified. - Stop the service.
- The
build.gradlefile configures the heap space to 2 GiB. - The
src/main/resources/application.yamlfile enables HTTP/2. - Time-out is 60s for both client and server.
Each line in src/main/resources/scenarios/scenarios.csv configures a test scenario which is performed first for Java Virtual Threads, then for WebFlux.
| scenario | k6Config | delayCallDepth | delayInMillis | connections | requestsPerSecond | warmupDurationInSeconds | testDurationInSeconds |
|---|---|---|---|---|---|---|---|
| 5k-vus-and-rps-get-time | get-time.js | 0 | 100 | 5000 | 5000 | 10 | 300 |
| 20k-vus-smooth-spike-get-movies] | k6-20k-vus-smooth-spike-get-movies].js | 0 | 100 | 20000 | 0 | 300 |
scenario: Name of scenario. Is printed on top of each diagram.k6Config: Name of the K6 Config File which is assumed to be in theconfigfolderdelayCallDepth: Depth of recursive HTTP call stack to$approach/epoch-millisendpoint prior to server-side delay.- Mimics calls to upstream services which allow for reuse of the current platform thread.
- For example, a value of
0means that the service waits for$delayInMillismilliseconds immediately upon receiving a request. - Otherwise, it calls the
$approach/epoch-milliswith${delayCallDepth - 1}. - This results in a recursive HTTP-request-based descent into the service, creating a call stack of depth
$delayCallDepth.
delayInMillis: Server-side delay of each request, in milliseconds. Mimics a delay such as invoking a DB which allow for reuse of the current platform thread.connections: Number of TCP connections, i.e. virtual users.requestsPerSecond: Number of requests per second across all connections. Left empty for scenarios where the number of requests per second is organically derived based on the number of connections, the request latency, and any explicit client-side delays.warmUpDurationInSeconds: Duration of the warm-up iteration before the actual test. Warm-up is skipped if0.testDurationInSeconds: Duration of the test iteration.
- Unless noted otherwise, all tests were conducted on this test environment.
- Preparation: The system was rebooted before each test and quieted down as much as possible. The baseline total CPU use before test start was 0.3%.
- Co-location: Test driver (k6) and server under test (Spring Boot microservice) were co-located on the same physical machine. The aim of this benchmark is not to achieve maximum absolute performance, but rather to compare different server-side approaches with each other. Considering that the test driver and the load it produced was identical for the combination of server-side approach and scenario, this co-location should not affect the validity of the test results.
- CPU: Intel Core i7-6700K at 4.00GHz with 4 cores (8 threads)
- Motherboard: Asus Z170-A
- RAM: 32 GiB DDR4 (2 x Corsair 16 GiB, 2133 MT/s)
- Network: Loopback interface
- Virtualization: None; bare metal desktop
- OS: Ubuntu 22.04.4 LTS
- Kernel: 5.15.86-051586-generic
- Java: Amazon Corretto JDK 21.0.3.9.1
- Spring Boot 3.2.5
The following charts show the results of each scenario, sorted by ascending scenario load.
Any lines in the client-side or error-side log files which contain the term error (case-insensitive) are preserved. You can find them in error log files, located in the results folder alongside the generated PNG files.
Any failed requests appear both in the latency chart as red dots, as well as in the RPS chart as part of a continuous orange
line. Additionally, they leave a trace in the $approach-latency.csv file, if preserved by running the benchmark with the -C option:
- A very small latency below 3ms indicates that the client failed to establish a TCP connection. Example from the latency CSV file:
1715728866471,0.000000,0,dial: i/o timeout,1211. Such requests are not considered when reporting minimum latency since this could obscure the minimum latency of successful requests. - A very large latency above 60s indicates a server-side time-out. Example from the latency CSV file:
1715728861008,60001.327066,0,request timeout,1050.
This scenario aims to maintain a steady number of 5k virtual users (VUs, i.e. TCP connections) as well as 5k requests per second (RPS) across all users for 5 minutes:
- Each user issues a request and then waits. This wait between consecutive requests is controlled by k6 in order to achieve the desired number of RPS.
- The server-side delay is 100ms.
- The server returns the current millis since the epoch.
Like the previous scenario, but the response body contains a JSON of movies.
For further details, please see the movies section.
Like the previous scenario, but 10 virtual users and requests per second.
Like the previous scenario, but mimics a request to an upstream service.
- On receiving an incoming HTTP request, the service calls itself via HTTP.
- This secondary request then waits 100 milliseconds.
This scenario ramps up virtual users (and thus TCP connections) from 0 to 25k in multiple steps, then back down:
- Each step has a short riser time when users are increased, followed by a longer tread time when users are held constant.
- Each user issues a request, waits for the response, and then waits for a random delay between 1s and 3s. This reduces the load and better mimics real user interactions with a service, assuming the service calls are driven by user interactions with a website that relies on the service under test.
- The server-side delay before returning a response is 100ms.
Like the previous scenario, but linear ramp-up and down.
Like the previous scenario, but instead of just getting movies, we are now additionally saving them:
- 75% of requests are GET requests which are split into three groups, each requesting movies by a different director.
- 25% of requests are POST requests.
For further details, please see the movies section.
Like the previous scenario, but mimics call to upstream service as explained in 10k-vus-and-rps-get-movies-call-depth-1.
Note
For loom-netty and webflux-netty, this scenario was CPU-contended on the test environment upon reaching ca. 5,000 RPS.
Whilst causing no errors, it drastically increased latencies.
Like 25k-vus-smooth-spike-get-post-movies, but scaling up to 60k users and executed within a VirtualBox VM on more powerful hardware, using a different Linux Kernel version. The rest of the setup is identical.
- CPU: Intel Core i7-12700K at 5GHz with 12 cores (20 threads)
- Motherboard: Asus ProArt Z690 Creator
- RAM: 64 GiB DDR4 (2 x Corsair 32 GiB, 5600MT/s)
- Network: Loopback interface
- Virtualization: VirtualBox 7.0.14 r161095 on bare metal desktop. All cores and 32GiB assigned to VM.
- Host OS: Windows 11 Pro (10.0.22631)
- Client OS: Ubuntu 22.04.4 LTS
- Client Kernel: 6.5.0-28-generic
- Java: Amazon Corretto JDK 21.0.2+13-LTS
- Spring Boot 3.2.5
