msbaz2013 / cs6301_sdn

Machine Learning Approach for an Anomaly Intrusion Detection System using ONOS

University of Texas at Dallas
CS 6301-503 Software Defined Networking

Professor:
      Timothy Culver

Team Members:
      Stephen Blystone
      Taniya Riar
      Juhi Bhandari
      Ishwank Singh

Project Name:
      Machine Learning Approach for an Anomaly Intrusion Detection System using ONOS

Project Report:
      Machine Learning Approach for an Anomaly IDS using ONOS.docx

Project Presentation:
      Project Presentation.pptx

======================================================================

SETTING UP THE PROJECT

For the ONOS VM:
      Follow the instructions in the "ONOS 1.12 installation Guide.docx" Guide.

For the Mininet VM:
      Follow the instructions in the "Mininet VM Guide.docx" Guide.

For the Application VM:
      Follow the instructions in the "App Installation Guide.docx" Guide.

======================================================================

RUNNING THE PROJECT

Follow the instructions to run the project and to stop running the project in the "Run Applications Guide.docx" Guide.

======================================================================

FILES IN PROJECT DIRECTORY

• Elasticsearchdata_csv.ipynb
  • iPython Jupyter Notebook used to visualize the ElasticsearchNormaldata.csv and use elbow graph to determine K value.


• ElasticsearchNormaldata.csv
  • "Normal" traffic data stored in Elasticsearch.


• Flow Diagram.pptx
  • Diagrams used in presentations.


• Important Netflow Fields.txt
  • Sample NetFlow data captured with only important fields remaining.


• Machine Learning Approach for an Anomaly IDS using ONOS.docx
  • Project Report


• Netflow Field Explanations.txt
  • Explanation of NetFlow fields.


• Project Presentation.pptx
  • Project Presentation.


• README.txt
  • This README file.


• sampleNetflowData.txt
  • Sample raw NetFlow data.


• SDN Project Proposal.docx
  • Our Project Proposal.

======================================================================

PROJECT GUIDES (located in the "Guides" directory)

• App Installation Guide.docx
  • How to install and configure everything in the Application VM.


• Mininet VM Guide.docx
  • Install required packages into the VM to run Mininet and trigger the anomalies.


• NetFlow Guide.docx
  • Guide for how to configure NetFlow on Open vSwitch.


• ONOS 1.12 installation Guide.docx
  • 3 methods of installing and configuring ONOS:
    1. Option 1 installs an OVA file and provides a link to a Distributed ONOS tutorial.
    2. Option 2 installs ONOS as a service (I did not get this to work).
    3. Option 3 is the recommended option. There is also information for configuring IntelliJ if building an Internal ONOS application.


• ONOS Rest API Guide.docx
  • Contains information on how to view a nice webpage on localhost (after launching ONOS) to query the ONOS REST API.


• Run Applications Guide.docx
  • How to start and stop all applications in the Big Data pipeline and run the demo.

======================================================================

CODE DESCRIPTIONS (located in the "Code/src" directory)

NOTE: All code files use UNIX EOL characters (line-feed "\n"). Opening these files in most Windows programs will not maintain the formatting, since Windows expects carriage-return line-feed "\r\n". If you use Windows, opening the files using Notepad++ will maintain the correct formatting.

• Client.py
  • Generates random “Normal” traffic.


• index_ES.txt
  • Information placed into Kibana Dev Tool to create our Elasticsearch index.


• scapyPortScan.py
  • Use Python library Scapy to perform a UDP port scan from port 1 to port 65535 on the target device.


• Server.py
  • Receives messages from Client.py from other hosts and responds.


• setup_topo.py
  • Setup Mininet topology, configure Open vSwitches with NetFlow, call Client.py and Server.py for each Mininet host.


• sparkKafka.py
  • Perform feature engineering to get our features and send to Elasticsearch.


• sparkMachineLearning.py
  • Train K-Means algorithm on data in Elasticsearch, perform feature engineering on new data, standardize new data and check if anomaly. If anomaly detected, send REST API call to ONOS.