Skip to content

Firewall False Data Injection IDS

This code is an implementation of a network anomaly detection system using packet size analysis. The purpose of this code is to load a training dataset consisting of legitimate network traffic captured in a “.pcap” file, calculate the mean and standard deviation of packet sizes in the dataset, and use this information to train an IsolationForest model to detect anomalies in live network traffic.

The code defines a function called “traffic_process” that extracts packet sizes from the live network traffic and compares the mean packet size of the captured traffic to the mean packet size of the training dataset. If the mean packet size of the captured traffic is significantly higher than that of the training dataset (i.e., exceeds two standard deviations), an alert is raised to indicate a potential anomaly. The code also excludes unauthorized protocols such as ESP and ISAKMP.

The code continuously captures live network traffic using the Scapy module’s “sniff” function and calls the “traffic_process” function to analyze the traffic. The script waits for two seconds before capturing another batch of network traffic.

This code can be used as a basis for an intrusion detection system to detect anomalies in network traffic, which can help to prevent cyber-attacks or unauthorized access to network resources.

from scapy.all import *
from sklearn.ensemble import IsolationForest
from sklearn.preprocessing import StandardScaler
import numpy

# Load dataset and extract features
dataset = rdpcap('datasets/training.pcap')
packetsizes = []
for packet in dataset:

# Calculate mean and standard deviation of packet sizes in dataset
mean_packetsize = numpy.mean(packetsizes)
std_packetsize = numpy.std(packetsizes)

# Define the IsolationForest model and the scaler for packet size normalization
isolationforest = IsolationForest(n_estimators=100, contamination=0.05)
scaler = StandardScaler()

# Train the scaler and fit the IsolationForest model on dataset, 1)), 1)))

def traffic_process(traffic):
    global mean_packetsize, std_packetsize, isolationforest, scaler
    # Extract packet sizes from network traffic
    traffic_packetsizes = []
    for packet in traffic:
        packet_size = len(packet)
        # exclude the unauthorized protocols
        if packet.haslayer(TCP) and not packet.haslayer(ESP) and not packet.haslayer(ISAKMP):
            print("ALERT: unauthorized protocol")

    # Calculate the mean packet size of the captured network traffic
    mean_traffic_packetsize = numpy.mean(traffic_packetsizes)
    # If the mean packet size of captured network traffic 
    # is significantly higher than that of dataset, print alert
    if mean_traffic_packetsize > mean_packetsize + 2*std_packetsize:
        print("ALERT: anomaly detected")
        print("Packet size OK")
    print("Valid average packet size: ", mean_packetsize)
    print("Captured average packet size:", mean_traffic_packetsize)

while True:
    capt_traff = sniff(count=5000)