A Framework for Anonymizing Network Flows using (k,l)-anonymization
- This project is in its alpha version. Further releases will have comments and detailed work flow.
- Redis (a key-value store) is used to map real IP addresses prefixes (first 16 bit) with fake prefixes that would be used for analysis purposes.
- The script expects a file (flows.csv) with the following format
SrcAddr,DstAddr,Proto,SrcBytes,DstByteswhere SrcAddr is the source IP address,DstAddr is the destination IP address,Proto is the protocol,SrcBytes is outgoing BPP (Bytes per Packet), and DstBytes is the ingoing BPP
SrcAddr,DstAddr,Proto,SrcBytes,DstBytes
10.0.2.15,10.0.2.255,udp,243,0
10.0.2.15,10.0.2.255,udp,216,0
10.0.2.15,10.0.2.255,udp,276,0
10.0.2.15,10.0.2.2,arp,42,64
10.0.2.15,186.50.139.76,udp,114,0
10.0.2.15,186.18.3.72,udp,114,0
10.0.2.15,82.231.139.51,udp,114,0
- Install Redis
- Clone the project:
git clone https://github.com/Lekssays/networkflowanony.git - Install requirements:
pip3 install -r requirements.txt - Change permission:
chmod +x preprocess.sh - Run the script:
./preprocess.sh flows.csv