Skip to main content
Join the official 2019 Python Developers SurveyStart the survey!

A machine learning/AI based approach to protecting your devices against ddos attacks

Project description

deepdos

Description

Welcome to deepdos, the python program written to monitor and potentially secure your network from ddos attacks! While not currently utilizing deep learning to classify packets, deepdos currently utilizes logistic regression in order to classify packets and has so far been trained on 200,000 packets from all sorts of DDOS attacks and normal traffic setup. This project couldn't have been done without the help of the Canadian Institute for Cybersecurity with providing both the original flow dataset and tool to create flow csvs from .pcap files. Their site and all resources have been linked at the bottom.

Goals

Short term goals

  • [ ] Add LR test metrics on startup
  • [ ] Update LR to use better data for better performance
  • [x] Add command line interface

Long term goals

  • [ ] Convert Logistic regression model to a neural network
  • [ ] Support both macos and Linux (potentially Windows as well if pcap is easy)
  • [ ] Add ddos mitigation/firewall rule support

How to run/setup

Running from scratch

deepdos is currently only available on linux, but can simply be run by these two commands:

# clone repo
git clone https://github.com/C3NZ/deepdos

# Install dependencies and setup the projects virtual environment
source bash/setup.sh

# Execute the script (Needs sudo in order to execute both tcpdump and iptables)
python3 main.py -h

This will load you into a virtualenv with all of the dependencies installed and ready to use.

To remove all of the dependencies after you're done using the tool, you can simply run:

source bash/remove.sh

and then remove the folder from your computer :)

This will immediately start creating necessary folders, capturing packets, and then identifying the traffic that is being exchanged in and out of your current computer.

This also assumes that you have java installed for the program to execute the CICFlowMeter jar file.

Installing with pip

Linux

sudo apt install libpcap-dev python3-dev python3-setuptools
pip3 install deepdos

Macos

brew install libpcap
pip3 install deepdos

Usage

usage: main.py [-h] [-i INTERFACE] [-n NAUGHTY_COUNT] [--find-interface]
               [--firewall FIREWALL] [--model-type MODEL_TYPE]

Welcome to deepdos, the machine learning/ai based ddos analysis/mitigation
service

optional arguments:
  -h, --help            show this help message and exit
  -i INTERFACE          [REQUIRES SUDO] The network interface for deepdos to
                        listen to (default: None)
  -n NAUGHTY_COUNT      The amount of malicious flows that can come from a
                        given address (default: 10)
  --find-interface      List all of your devices network interfaces. Good if
                        you don't know what interfaces your device has
                        (default: False)
  --firewall FIREWALL   [REQUIRES SUDO] Turn on firewall mode for the given
                        system. linux for Linux systems and macos for mac (Not
                        yet supported) (default: None)
  --model-type MODEL_TYPE
                        The model that you would like to use for classifying
                        the data (default: lr-stable-0.9.0.pickle)
usage: src [-h] [-i INTERFACE] [-n NAUGHTY_COUNT] [--find-interface]
           [--firewall FIREWALL] [--model-type MODEL_TYPE]

How to deploy

You can deploy this on your own machine, but production use will come in the future.

Live deployments

This will be on pypi soon :)

How to contribute

Fork the current repository and then make the changes that you'd like to said fork. Upon adding features, fixing bugs, or whatever modifications you've made to the project, issue a pull request to this repository containing the changes that you've made and I will evaluate them before taking further action. This process may take anywhere from 3-7 days depending on the scope of the changes made, my schedule, and any other variable factors.

Resources

UNB datasets

CICnetflowmeter

CIC License

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for deepdos, version 0.9.95
Filename, size File type Python version Upload date Hashes
Filename, size deepdos-0.9.95-py3-none-any.whl (16.4 MB) File type Wheel Python version py3 Upload date Hashes View hashes
Filename, size deepdos-0.9.95.tar.gz (16.3 MB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page