Open source controller for a secure Software Defined Network
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
attack_scripts
modules extra file removed Oct 29, 2014
.gitignore Changed flood to identity in arp spoofing. Reordered the modules in t… Oct 31, 2014
LICENSE merge fixed Oct 29, 2014
README.md updated readme Oct 31, 2014

README.md

Galatine

Galatine is a project for detecting layer 2 attacks and automatically remediating them from within an SDN controller. It is written in the Pyretic language.

For information on the Pyretic project visit www.frenetic-lang.org/pyretic. It is recommended that you run Pyretic in the Pyretic VM, which can be found at that website.

In the modules/ directory, there are several elements of a Pyretic controller that can be combined to add various security aspects to the network. The main module is galatine_controller.py. To run this, you must add the modules in this directory to the pyretic/modules directory in the pyretic repo (another option is to create a symbolic link). Once that is done, navigate to the pyretic directory and execute the command:

pyretic.py -m r0 pyretic.modules.galatine_controller

Config

By default, galatine provides protection against the following:

  • ARP spoofing
  • MAC flooding
  • DoSing the controller
  • Ethernet control frames

To configure this, edit the security_measures variable in galatine_controller.py (simply comment out any protections you do not wish to use).

By default, galatine only provides one option for remediation, namely cutting off the offending host(s) from the network. In future versions there will be more than one option, and remediation strategy will be configurable.

Environment

The easiest environment in which to experiment with this controller is mininet. Recommended command:

sudo mn --controller remote --topo single,4

Please note that mininet does not work with the promiscuous mode detection. This is due to the fact that mininet hosts are always in promiscuous mode.

Testing

You can run the various scripts in attack_scripts to test out the controller. The example mininet command would be:

h1 arp_reply_spoof.py -v h2 -r h3

This will attempt to establish a MitM attack from h1 between h2 and h3.