Statitical Anomaly Detector of Internet Traffic (SADIT)
The repo of SADIT has been mirgrated to Github.com. Please visit https://github.com/hbhzwj/SADIT
What is it?
SADIT aims to provide research community an easy-to-use tool to validate and test statistical anomaly detecting method in both simulated and real environment.
In the current developing stage, we focus on simulation. SADIT uses fs(flow-max) nework simulator, an efficient & light-weight network simulator developed by UW Madison, to simulate the network flow traffic. For more information about fs(flowmax) simulator, visit http://cs.colgate.edu/~jsommers/#code Also refer paper "Efficient Network-wide Flow Record Generation" Infocom 2011.
SADIT now has the following three modules: ** Configure ** fs-simulator can only parse DOT file. this module will generate different types of DOTS file according the scenaios specfied in settings.py
** Simulator ** this is a revised fs simulator which supports: 1. export abnormal flows record
** Detector ** Implements anomaly detection algorithm in paper "STATICAL APPROACHES TO DATA EXFILTRATION DETECTION" submitted to CDC2012.
please type the ./cmdsadit and help documents will appear
You need to specify the environment variable SADIT_ROOT before running it.
this software depends on all softwares that fs-simulate depends on: - ipaddr (2.1.1)
- networkx (1.0) Get_<http://networkx.lanl.gov/download/networkx/networkx-1.0.1.tar.gz>
- pydot (1.0.2) Get_<http://pydot.googlecode.com/files/pydot-1.0.2.tar.gz>
- pyparsing (1.5.2) Get_<http://downloads.sourceforge.net/project/pyparsing/pyparsing/pyparsing-1.5.2/pyparsing-1.5.2.tar.gz?r=http%3A%2F%2Fsourceforge.net%2Fprojects%2Fpyparsing%2Ffiles%2Fpyparsing%2Fpyparsing-1.5.2%2F&ts=1332828745&use_mirror=softlayer>
- py-radix (0.5) Get_<http://py-radix.googlecode.com/files/py-radix-0.5.tar.gz>
besides: it requires python-dev numpy matplotlib
if you are in debain base system. you can simple use sudo apt-get install python-dev sudo apt-get install python-numpy sudo apt-get install python-matplotlib
Please see the file called LICENSE.