The purpose of this project is to evaluate serialization and deserialization performance of JSON libraries in Java. The following libraries are evaluated:
- jackson
- genson
- fastjson
- gson
- org.json
- javax-json (from Oracle)
- json-io
- flexjson
- boon
- json-smart
- johnzon
- logansquare
- dsl-json
- json-simple
- nanojson
- moshi
This benchmark tests throughput performance of serialization and deserialization algorithms of the databind and stream API when available. Random payloads of various sizes are generated at runtime before each benchmark.
Four different sizes of payloads are evaluated in the charts below: 1 KB, 10 KB, 100 KB and 1 MB. And it is possible to generate on the fly any size of payloads.
Each benchmark has been written to read bytes from RAM and write to reusable output streams in RAM when possible, strings are rarely generated. All data is randomly generated upon static loading.
This benchmark does NOT evaluate:
- compression performance or efficiency
- payloads bigger than 1.1 MB (would be easy to do though)
- RAM utilization
The benchmarks are written with JMH and for Java 8.
The results here-below were computed on November the 21st, 2016 with the following libraries and versions:
| Library | Version |
|---|---|
| jackson | 2.8.4 |
| genson | 1.4 |
| fastjson | 1.2.20 |
| gson | 2.8.0 |
| org.json | 20090211 |
| javax-json | 1.0, 1.0.4 |
| json-io | 4.9.0 |
| flexjson | 3.3 |
| boon | 0.34 |
| json-smart | 2.2.1 |
| johnzon | 0.9.5 |
| logansquare | 1.3.7 |
| dsl-json | 1.1.2 |
| simplejson | 1.1.1 |
| nanojson | 1.2 |
| jodd json | 3.8.0 |
| moshi | 1.3.1 |
All graphs and sheets are available in this google doc.
# JMH 1.15 (released 52 days ago)
# VM version: JDK 1.8.0_45, VM 25.45-b02
# VM invoker: /Library/Java/JavaVirtualMachines/jdk1.8.0_45.jdk/Contents/Home/jre/bin/java
# VM options: -XX:+AggressiveOpts -Xms2G -Xmx2G
# Warmup: 3 iterations, 1 s each
# Measurement: 5 iterations, 2 s each
# Timeout: 10 min per iteration
# Threads: 16 threads, will synchronize iterations
# Benchmark mode: Throughput, ops/time
Model Name: MacBook Pro
Processor Name: Intel Core i7
Processor Speed: 2.6 GHz
Number of Processors: 1
Total Number of Cores: 4
L2 Cache (per Core): 256 KB
L3 Cache: 6 MB
Memory: 16 GB
By default, running ./bench ser (./bench deser respectively) will run
all -- stream and databind -- serialization (deserialization respectively)
benchmarks with 1 KB payloads of Users.
You can also specificy which libs, apis, payload-sizes and number of iterations (and more) you want to run. For example:
./bench deser --apis stream --libs genson,jackson
./bench ser --apis databind,stream --libs jackson
./bench deser --apis stream --libs dsljson,jackson --size 10 --datatype users
Type ./bench help ser or ./bench help deser to print help for those
commands.
If you wish to run all benchmarks used to generate the reports above,
you can run ./bench-all. This will take several hours to complete, so
be patient.
Any help to improve the existing benchmarks or write ones for other libraries is welcome.
Adding a JSON library to the benchmark requires little work and you can find numerous examples in the commit history. For instance:
- Addition of moshi: https://github.com/fabienrenaud/java-json-benchmark/commit/6af2c0a7091b12a9dc768e49499682b97ea57ff6
- Addition of jodd: https://github.com/fabienrenaud/java-json-benchmark/commit/288a4e61496588ed4c0a80e1f107f34f9a2c985c
- Addition of json-simple: https://github.com/fabienrenaud/java-json-benchmark/commit/1e1e559c39a6eddc3dd7d7cea777fc7861415469
Pull requests are welcome.