How to calculate Shannon entropy based on HTTP header

Question

How to calculate Shannon entropy based on HTTP header

Asked 6 years, 10 months ago

Viewed 597 times

9

Shannon’s entropy is given by the formula:

Where Ti will be the data extracted from my network dump (dump.pcap).

The end of an HTTP header on a normal connection is marked by \r\n\r\n:

Example of an incomplete HTTP header (could be a denial of service attack):

My goal is to calculate the entropy of the number of packets with \r\n\r\n and without \r\n\r\n in order to compare them.

I can read the PCAP file like this:

import pyshark

pkts = pyshark.FileCapture('dump.pcap')

The entropy based on the IP numbers I made:

import numpy as np
import collections

sample_ips = [
    "131.084.001.031",
    "131.084.001.031",
    "131.284.001.031",
    "131.284.001.031",
    "131.284.001.000",
]

C = collections.Counter(sample_ips)
counts = np.array(list(C.values()),dtype=float)
#counts  = np.array(C.values(),dtype=float)
prob    = counts/counts.sum()
shannon_entropy = (-prob*np.log2(prob)).sum()
print (shannon_entropy)

Any idea? It is possible/makes sense to calculate entropy based on the number of packages with \r\n\r\n and without \r\n\r\n? Or is it something that doesn’t make sense? Any idea how to calculate?

The network dump is here: https://ufile.io/y5c7k

Some lines from it:

30  2017/246 11:20:00.304515    192.168.1.18    192.168.1.216   HTTP    339 GET / HTTP/1.1 


GET / HTTP/1.1
Host: 192.168.1.216
accept-language: en-US,en;q=0.5
accept-encoding: gzip, deflate
accept: */*
user-agent: Mozilla/5.0 (X11; Linux i686; rv:45.0) Gecko/20100101 Firefox/45.0
Connection: keep-alive
content-type: application/x-www-form-urlencoded; charset=UTF-8

1 answer

Browser other questions tagged python python-3.x python-2.7 security-guard

You are not signed in. Login or sign up in order to post.

by Begnini • **1,681** points · Answer 1 · 2017-09-06T01:03:06+00:00

I don’t know what the structure of your package returned by pyshark looks like, but I imagine it has 2 information, the IP address and the contents of the package. Imagining that you have these 2 information in a Dict, you could do something like:

pkgs = [
    {
        'ip': '127.0.0.1',
        'content': 'Im a http header\r\n\r\n<html><body>',
    },
    {
        'ip': '127.0.0.1',
        'content': 'Im a not a http header',
    },
    {
        'ip': '127.0.0.2',
        'content': 'Im a http header\r\n\r\n<html><body>',
    },
    {
        'ip': '127.0.0.2',
        'content': 'Im a not a http header',
    },
    {
        'ip': '127.0.0.2',
        'content': 'Im a not a http header too',
    }
]

def is_http(content):
    return '\r\n\r\n' in content

classified_pkgs = [(p['ip'], is_http(p['content'])) for p in pkgs]
>> [('127.0.0.1', True),
>> ('127.0.0.1', False),
>> ('127.0.0.2', True),
>> ('127.0.0.2', False),
>> ('127.0.0.2', False)]

Then simply calculate the probabilities as you calculated before:

import numpy as np
import collections

counter = collections.Counter(classified_pkgs)
counts  = np.array(list(counter.values()),dtype=float)

prob = counts/counts.sum()
shannon_entropy = (-prob * np.log2(prob)).sum()
print (shannon_entropy)