Skip to content
This repository has been archived by the owner on Jul 31, 2019. It is now read-only.

Benchmarks

Daniel Oltmanns edited this page Nov 16, 2017 · 2 revisions

Welcome to the zap wiki!

Benchmarks

This is where the detailed benchmark results live, and we describe numbers in more detail.

Iron

Running 10s test @ http://127.0.0.1:8080
  16 threads and 500 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   214.85us  622.27us  78.67ms   99.24%
    Req/Sec    77.20k    51.31k  191.91k    51.98%
  Latency Distribution
     50%  183.00us
     75%  219.00us
     90%  304.00us
     99%  746.00us
  3103506 requests in 10.09s, 337.41MB read
Requests/sec: 307581.17
Transfer/sec:     33.44MB

What do we see?

Iron has a quick response time, and handles about ~307k requests per second.

Zap

Running 10s test @ http://127.0.0.1:8080
  16 threads and 500 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency     1.08ms    1.90ms  85.45ms   89.78%
    Req/Sec    56.88k    18.89k  126.22k    71.86%
  Latency Distribution
     50%  327.00us
     75%    1.42ms
     90%    3.02ms
     99%    7.60ms
  9079097 requests in 10.08s, 441.58MB read
Requests/sec: 912832.31
Transfer/sec:     40.90MB

What do we see?

Zap has an average but still fast response time, and handles about ~900k requests per second.

What's the difference

Iron can answer the requests in the benchmark quicker, but only handles ~1/3 of the requests zap is doing in the same time. Zap has a higher response time, because it can't serve that many request at the same time with in the same latency.

Clone this wiki locally