CLOAK

Anonymous IP allocation through Heroku Worker Dynos.

BACKGROUND

"Cloak" is call based API which means that it is controlled by requesting variations of a URL. It is broken into two running processes, a managing script on an AWS instance, and a series of pre-configured heroku Dynos which each have a built in scraping module. Each dyno requests a dynamic IP address and performs the data call for you. It supports http and https, Javascript, Jquery, AJAX and anything else with a GET request. If this is something you end up using I will add in the POST functionality for when you might need to login.

The scripts are included as a proof of concept around the app currently hosted at: http://cloak.herokuapp.com. Cloak is written in Node 0.10.8 but all features are supported across all major languages (Sorry, lisp and fortran are off the table).

Heroku (www.heroku.com) is a platform that offers VM and hosting services for applications. Often touted as a weakness of Heroku, the platform rotates the IP address of each app deployment. On the other hand, the dynamic assignments are incredibly valuable if each dyno were to act as a proxy for requests from a local machine.

USE CASE

For services where a "fresh" IP address is needed, Cloak allows you to access data anonymously from a new and anonymous IP without the need for proxy services or "spoofing".

To use Cloak, you append every url call to the Cloak address. 1- ‘http://cloak.herokuapp.com/?http://www.whatismyip.com’ the response is then returned as a complete browser object.

2- Cloak supports https secure connections automatically. Simply ensure the calls you want are the complete https url and Cloak will adapt.

Now your script has it’s own unique URL assigned to outgoing requests.

What to do with throttling/blocked IPs/blacklisting? Simply include a latency timer and an if statement in your script. If the latency gets too long, load “http://cloak.herokuapp.com/refresh” and wait for 15 seconds for the dynos to boot down and reconnect with Heroku and Route Director.

And there is a secret function :)

For large scale scrapes, pausing 15 seconds is both annoying and can start to add up. Especially as Google begins throttling at 250 requests. By calling, “http://cloak.herokuapp.com/cluster” you will tell the scalling manager to boot 100 worker threads each capapable of simaltaneously scrapping your requests. Each one works in a cascading error fashion so that as one IP gets blocked or throttled it kills the dyno and works with the next. ONLY use this were approved because it requires we pay Heroku for some of the processing time (using the blocking version of Cloak is free for us).

For example, this script (http://github.com/thnkr/cloak/examples/basic.py) demos controlling the cloak and scrapes whatismyip.com as an example so you can see the change.

import time
import codecs
import re
import urllib2
import pprint
from bs4 import BeautifulSoup, SoupStrainer
import string
import sys
freshIP = 'http://cloak.herokuapp.com/refresh'

target_url = 'http://www.whatismyip.com/'
header =  {'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.11 (KHTML, like Gecko) Chrome/23.0.1271.64 Safari/537.11','Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8','Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.3','Accept-Encoding': 'none','Accept-Language':     'en-US,en;q=0.8','Connection': 'keep-alive'}
open_url = urllib2.Request(target_url, headers=header)
target_data = urllib2.urlopen(open_url)
soup = BeautifulSoup(target_data)
links = soup.findAll('div',{'id':'greenip'})[0].getText()
print ' ' 
print target_url
print 'Publicly...'
print links

target_url = 'http://cloak.herokuapp.com/?http://www.whatismyip.com/'
header =  {'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.11 (KHTML, like Gecko) Chrome/23.0.1271.64 Safari/537.11','Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8','Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.3','Accept-Encoding': 'none','Accept-Language': 'en-US,en;q=0.8','Connection': 'keep-alive'}
open_url = urllib2.Request(target_url, headers=header)
target_data = urllib2.urlopen(open_url)
soup = BeautifulSoup(target_data)
links = soup.findAll('div',{'id':'greenip'})[0].getText()
print 'http://wwww.whatismyip.com/'
print 'In a cloak...'
print links
time.sleep(2) # "Paw"se for effect. 

print 'Now lets change cloaks...\nCALLED URL: cloak.herokuapp.com/refresh\nWAITING 15 Seconds'
open_url = urllib2.Request(freshIP, headers=header) # Note the freshIP variable. 
target_data = urllib2.urlopen(open_url)
soup = BeautifulSoup(target_data)
time.sleep(15)

print '\nUse New IP:'
target_url = 'http://cloak.herokuapp.com/?http://www.whatismyip.com/'
header =  {'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.11 (KHTML, like Gecko) Chrome/23.0.1271.64 Safari/537.11','Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8','Accept-Charset':   'ISO-8859-1,utf-8;q=0.7,*;q=0.3','Accept-Encoding': 'none','Accept-Language':'en-US,en;q=0.8','Connection': 'keep-alive'}
open_url = urllib2.Request(target_url, headers=header)
target_data = urllib2.urlopen(open_url)
soup = BeautifulSoup(target_data)
links = soup.findAll('div',{'id':'greenip'})[0].getText()
print 'http://wwww.whatismyip.com/'
print 'Our fresh IP:'
print links

The output of the example script looks like this:

http://www.whatismyip.com/
Publicly...
54.225.220.147

http://wwww.whatismyip.com/
In a cloak...
184.72.209.15

Now lets change cloaks...
CALLED URL: cloak.herokuapp.com/refresh
WAITING 15 Seconds

Our fresh IP:
http://wwww.whatismyip.com/
In a new cloak...
54.224.225.80

INSTALLATION

• Create a free Heroku account at www.Heroku.com.

• Install the Heroku Toolbelt (https://toolbelt.heroku.com/)

• Download the repo and install the necessary packages.

    git clone git://github.com/thnkr/cloak.git`
    cd cloak/app
    git init
    npm install
    heroku create
    git add .
    git commit -m "Some commit message."
    git push heroku master
  

• Navigate to your hostaddress on port 3000. If this is on AWS make sure you have opened the security group on the port.

• ERRORS Were you unabnle to authenticate? Follow these steps: https://devcenter.heroku.com/articles/keys

Make sure you change the Child_Process directory to match your Heroku app. This can be found in cloak/front-end/front-end.js.

Cloak is written in Node JS and spawns a Child Process to interact with Heroku. Read more about these here: http://nodejs.org/api/child_process.html.

LINKS

• icodeforlove/node-requester
• https://devcenter.heroku.com/articles/third-party-buildpacks