ddos-mlp-mitigation

git clone git://git.codymlewis.com/ddos-mlp-mitigation.git
Log | Files | Refs | Submodules | README | LICENSE

README.md (1116B)


      1 # DDoS MLP Mitigator
      2 A combination of SDN and a multi-layer perceptron neural network that results in the
      3 mitigation of DDoS attacks
      4 
      5 ## Requirements
      6 - python3
      7 - pip
      8 - rust
      9 - cargo
     10 
     11 ## Installation
     12 ```
     13 setup.sh
     14 ```
     15 
     16 ## Generating data
     17 First start the controller in generate data mode:
     18 ```
     19 ./network_controller.py --gen-data
     20 ```
     21 
     22 Then start the network in normal interactions training mode (this uses mininet
     23 so it will probably require root privileges to run):
     24 ```
     25 ./create_network --normal
     26 ```
     27 
     28 Once done, train for the attack state. Start the controller in generate attack
     29 data mode:
     30 ```
     31 ./network_controller.py --attack --gen-data
     32 ```
     33 
     34 Then start the network in attack interactions training mode:
     35 ```
     36 ./create_network --all-attack
     37 ```
     38 
     39 ## Training the MLP
     40 Simply run the following:
     41 ```
     42 ./network_controller.py --train
     43 ```
     44 
     45 ## Run DDoS Mitigation
     46 Start the controller in detection mode:
     47 ```
     48 ./network_controller.py --detect
     49 ```
     50 
     51 Then start the network in attack and CLI mode:
     52 ```
     53 ./create_network --attack --cli
     54 ```
     55 
     56 The user should be able to ping the attack target with the following command:
     57 ```
     58 u0 ping t0
     59 ```