Accurate proof that Ruby is closer to the bear metal than Node.js

bench_node.sh

#!/usr/bin/env bash

siege -c 10 -b -t 30s http://localhost:9393

bench_ruby.sh

#!/usr/bin/env bash

siege -c 10 -b -t 30s http://localhost:9292

config.ru

require 'rack'

class Application
  def call(env)
    return [200, {'Content-Type' => 'text/plain'}, ['Hello world']]
  end
end

run Application.new

node.js

var http   = require('http');
var server = http.createServer(function(request, response)
{
    response.writeHead(200, {'Content-Type': 'text/plain'});
    response.end('Hello world');
});

server.listen(9393, '0.0.0.0');

results.txt

ruby 1.9.3p448 (2013-06-27 revision 41675) [x86_64-linux]:

** SIEGE 3.0.5
** Preparing 10 concurrent users for battle.
The server is now under siege...
Lifting the server siege...      done.

Transactions:                 158092 hits
Availability:                 100.00 %
Elapsed time:                  29.06 secs
Data transferred:               1.66 MB
Response time:                  0.00 secs
Transaction rate:            5440.19 trans/sec
Throughput:                     0.06 MB/sec
Concurrency:                    9.90
Successful transactions:      158092
Failed transactions:               0
Longest transaction:            0.02
Shortest transaction:           0.00

ruby 2.1.1p76 (2014-02-24 revision 45161) [x86_64-linux]:

** SIEGE 3.0.5                                    
** Preparing 10 concurrent users for battle.      
The server is now under siege...                  
Lifting the server siege...      done.            
                                                  
Transactions:                 222639 hits         
Availability:                 100.00 %            
Elapsed time:                  29.91 secs         
Data transferred:               2.34 MB           
Response time:                  0.00 secs         
Transaction rate:            7443.63 trans/sec    
Throughput:                     0.08 MB/sec       
Concurrency:                    9.88              
Successful transactions:      222639              
Failed transactions:               0              
Longest transaction:            0.02              
Shortest transaction:           0.00              

rubinius 2.2.6.n73 (2.1.0 f3c12adf 2014-03-14 JI) [x86_64-linux-gnu]:

** SIEGE 3.0.5
** Preparing 10 concurrent users for battle.
The server is now under siege...
Lifting the server siege...      done.

Transactions:                 103199 hits
Availability:                 100.00 %
Elapsed time:                  29.28 secs
Data transferred:               1.08 MB
Response time:                  0.00 secs
Transaction rate:            3524.56 trans/sec
Throughput:                     0.04 MB/sec
Concurrency:                    9.94
Successful transactions:      103199
Failed transactions:               0
Longest transaction:            0.11
Shortest transaction:           0.00

jruby 1.7.10 (1.9.3p392) 2014-01-09 c4ecd6b on OpenJDK 64-Bit Server VM 1.7.0_51-b31 [linux-amd64]:

** SIEGE 3.0.5                                
** Preparing 10 concurrent users for battle.  
The server is now under siege...              
Lifting the server siege...      done.        
                                              
Transactions:                 180525 hits     
Availability:                 100.00 %        
Elapsed time:                  29.82 secs     
Data transferred:               1.89 MB       
Response time:                  0.00 secs     
Transaction rate:            6053.82 trans/sec
Throughput:                     0.06 MB/sec   
Concurrency:                    9.90          
Successful transactions:      180525          
Failed transactions:               0          
Longest transaction:            0.06          
Shortest transaction:           0.00          

Node.js v0.10.26:

** SIEGE 3.0.5
** Preparing 10 concurrent users for battle.
The server is now under siege...
Lifting the server siege...      done.

Transactions:                 190550 hits
Availability:                 100.00 %
Elapsed time:                  29.91 secs
Data transferred:               2.00 MB
Response time:                  0.00 secs
Transaction rate:            6370.78 trans/sec
Throughput:                     0.07 MB/sec
Concurrency:                    9.92
Successful transactions:      190550
Failed transactions:               0
Longest transaction:            0.02
Shortest transaction:           0.00

start_node.sh

#!/usr/bin/env bash

echo 'Node version:'

node --version

node node.js

start_ruby.sh

#!/usr/bin/env bash

echo 'Ruby version:'

ruby --version

export RACK_ENV=none

puma -e none -p 9292 -t 64:128

I feel like you're essentially recreating the plaintext test of the techempower benchmarks but with less warmup time and a slower server choice for JRuby. Here's a link to their latest published results, filtered to show only nodejs, rack-jruby, and rack-ruby.

http://www.techempower.com/benchmarks/#section=data-r8&hw=i7&test=plaintext&f=g0-0-0

In case you don't follow the link or it breaks: rack-jruby 269k req/s, nodejs 80k req/s, and rack-ruby 69k req/s.

The JSON parsing results at http://www.techempower.com/benchmarks/#section=data-r8&hw=i7&test=json&f=g0-0-0 are similar, with rack-jruby at 186k req/s, nodejs at 71k req/s, and rack-ruby at 61k req/s.

They don't have Rubinius in their test suite, but would probably be open to adding it for future rounds. You can try to recreate the jruby/ruby/node results by replacing puma with the 'torqbox' gem for jruby and unicorn behind nginx for ruby.