https://github.com/cannikin/performance

Testing Rails vs. Sinatra vs. Rack
https://github.com/cannikin/performance

Last synced: about 2 months ago
JSON representation

Testing Rails vs. Sinatra vs. Rack

• Host: GitHub
• URL: https://github.com/cannikin/performance
• Owner: cannikin
• Created: 2009-07-09T01:12:19.000Z (over 15 years ago)
• Default Branch: master
• Last Pushed: 2009-07-13T16:07:46.000Z (over 15 years ago)
• Last Synced: 2024-10-12T15:22:45.616Z (3 months ago)
• Language: Ruby
• Homepage:
• Size: 2.46 MB
• Stars: 1
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.rdoc

Awesome Lists containing this project

README

        
Here is the server config scripts I used to setup the Amazon EC2 instance.

These tests were run on a Medium-sized, High CPU Instance of EC2 running a bare bones

Ubuntu 8.10 install: ami-0372946a

You can copy the "EC2 Instance Config" section into a script and run it right on the server

to get everything set up.

= Test Procedure

After all packages were install and various servers were started and found to work

the instance was rebooted to start fresh. Apache and nginx were stopped:

  /etc/init.d/nginx stop

  /etc/init.d/apache2 stop

When switching to a new framework/server a test was run once to make sure everything

was loaded into memory before recording any results.

The following script was run with slight variations for each server type. It rans each test 

11 times (the first was discarded to make sure everything was loaded into memory and ready 

to go at full speed). Only the requests per second number for each test was returned to the 

command line:

  ab -n 10000 -c 50 -q http://127.0.0.1:3000/ | grep 'Requests per' && \

  ab -n 10000 -c 50 -q http://127.0.0.1:3000/ | grep 'Requests per' && \

  ab -n 10000 -c 50 -q http://127.0.0.1:3000/ | grep 'Requests per' && \

  ab -n 10000 -c 50 -q http://127.0.0.1:3000/ | grep 'Requests per' && \

  ab -n 10000 -c 50 -q http://127.0.0.1:3000/ | grep 'Requests per' && \

  ab -n 10000 -c 50 -q http://127.0.0.1:3000/ | grep 'Requests per' && \

  ab -n 10000 -c 50 -q http://127.0.0.1:3000/ | grep 'Requests per' && \

  ab -n 10000 -c 50 -q http://127.0.0.1:3000/ | grep 'Requests per' && \

  ab -n 10000 -c 50 -q http://127.0.0.1:3000/ | grep 'Requests per' && \

  ab -n 10000 -c 50 -q http://127.0.0.1:3000/ | grep 'Requests per' && \

  ab -n 10000 -c 50 -q http://127.0.0.1:3000/ | grep 'Requests per'

== Servers

Each server was the only one running at the time of the tests. Two ssh sessions

were running. One to start the server and the other to run Apache Bench.

=== Rack

  ruby /mnt/performace/rack/erb.rb

  ab -n 1000 -c 10 -q http://127.0.0.1:3000/

=== Sinatra

  ruby /mnt/performance/sinatra/erb.rb -e production -p 3000

  ab -n 1000 -c 10 -q http://127.0.0.1:3000/

=== Rails (served by Thin)

  thin start -c /mnt/performance/rails -e production -p 3000

  ab -n 1000 -c 10 -q http://127.0.0.1:3000/go/erb

=== Rails (served by Passenger)

  /etc/init.d/apache2 start

  ab -n 1000 -c 10 -q http://127.0.0.1:3000/go/erb

=== Nginx

  /etc/init.d/nginx start

  ab -n 1000 -c 10 -q http://127.0.0.1/

=== Apache

  /etc/init.d/apache2 start

  ab -n 1000 -c 10 -q http://127.0.0.1/