Featured domain

Backend services

HTTP services backed by PostgreSQL, each running in the same small, resource-limited container.

15 implementations

Backend services: implementations ranked by Throughput index
Rank by Throughput indexImplementation
1TypeScript · BunBunBare99.21.27ms75.9MiB200ms
2Rust · axumTokioMicro91.60.93ms21.6MiB173ms
3C# · ASP.NET Core 10 LTSJIT.NET 10Micro73.11.34ms144.1MiB501ms
4C# · ASP.NET Core 10 LTSNative AOT.NET 10Micro65.41.50ms93.2MiB690ms
5C# · ASP.NET Core 11JIT.NET 11Micro59.21.70ms155.6MiB496ms
6TypeScript · FastifyNode.jsMicro56.81.90ms223.5MiB582ms
7Go · net/httpGoBare55.31.86ms41.9MiB170ms
8C# · ASP.NET Core 11Native AOT.NET 11Micro52.61.78ms97.2MiB215ms
9Java · Spring BootJVMJVM (HotSpot)Full-stack45.66.95ms680.0MiB2,368ms
10Java · Spring BootJVM + AOT cacheJVM (HotSpot)Full-stack45.42.83ms671.4MiB1,486ms
11TypeScript · ExpressNode.jsMicro41.72.53ms220.0MiB532ms
12Elixir · Plug + BanditBEAM (Erlang/OTP 29)Micro32.44.87ms156.6MiB1,203ms
13PHP · FrankenPHPFrankenPHP (worker mode)Bare25.57.78ms134.8MiB307ms
14Java · Spring BootNative imageGraalVM Native ImageFull-stack17.038.56ms244.3MiB274ms
15Python · FastAPICPythonFramework14.114.19ms141.6MiB1,274ms

Fastest in each workload

Throughput · the best valid result in each of the 6 workloads behind the throughput index

Throughput index
Geometric mean over six workloads, each relative to the fastest in that workload. 100 means fastest in all six.
Latency under load
p99 at a fixed 5,000 requests per second.
Memory
Highest proportional set size (PSS) seen in any scenario.
Cold start
Container start until the first successful response.

Community

Run by its contributors

Implementations, domains and measurements all arrive as pull requests. Run the suite on your own machine and your results are published next to everyone else's, with the hardware they were measured on.

Published runs
2
Machines
1
Reference runs
2
Community runs
0

No community runs yet. Add the first. Reference runs are measured by the maintainers; every other merged run is labelled community.

Submit a run

Methodology

Measured the same way, for everyone

Numbers are only worth comparing when the conditions are identical. Every implementation gets the same box, the same work and the same statistics.

Read the methodology

Same limits

Each service runs in its own container limited to 2 CPUs and 1 GiB of memory, like a small production pod.

Isolated cores

Service, database and load generator are pinned to separate cores, so they never compete for CPU time.

  1. CPU 0: Service under test
  2. CPU 1: Service under test
  3. CPU 2: Environment (database)
  4. CPU 3: Environment (database)
  5. CPU 4: Load generator
  6. CPU 5: Load generator
  7. CPU 6: Kept idle
  8. CPU 7: Kept idle
  9. CPU 8: Environment (database)
  10. CPU 9: Environment (database)
  11. CPU 10: Load generator
  12. CPU 11: Load generator
  • Service under test 0, 1
  • Environment (database) 2, 8, 3, 9
  • Load generator 4, 10, 5, 11
  • Kept idle 6, 7

SMT is on. The service gets one logical CPU per physical core; the SMT siblings of those cores stay idle.

Repeated and robust

Several passes in rotating order. Results are medians with a 95 % bootstrap confidence interval; outliers are flagged, never silently removed.

Verified work

Every implementation must pass functional checks before it is measured, and the backend harness counts the database statements each request really runs.