Login Attempt Simulator

From the book Hands-On Data Analysis with Pandas.

Simulation of regular login activity on a site and random activity from hackers using brute-force password guessing attacks. The login process involves a username and password (no additional validation).

Assumptions

The simulator makes the following assumptions about valid users of the website:

Valid users come according to a Poisson process with an hourly rate that depends on the day of the week and the time of day. A Poisson process models arrivals per unit time (hour here) as a Poisson distribution with mean λ (lambda) and the interarrival times are exponential distributed with mean 1/λ.
Valid users connect from 1-3 IP addresses (unique identifier for devices using the Internet), which are 4 random integers in [0, 255] separated by periods. It is possible, although highly unlikely, that two valid users have the same IP address.
Valid users are unlikely to make many mistakes entering their credentials.

The simulator makes the following assumptions about the hackers:

The hackers try to avoid an account lockout by only testing a few username-password combinations rather than a full-blown dictionary attack (trying every password the hacker has in a dictionary of possible passwords that they maintain on every user). However, they don't add delays between their attempts.
Since the hackers don't want to cause a denial of service, they limit the volume of their attacks and only make one attempt at a time.
The hackers know the amount of accounts that exist in the system and have a good idea the format the usernames are in, but are guessing what they are exactly. They will choose to try to guess all 133 usernames or some subset of it.
Each attack is standalone, meaning there is a single hacker acting for each attack.
The hackers don't share information about which username-password combinations are correct.
The attacks come randomly.
Each hacker will use a single IP address, which is generated in the same way the valid user ones are. However, our simulator is capable of varying this IP address when vary_ips=True is passed to simulate().
Although highly unlikely, it is possible the hacker has the same IP address as a valid user. The hacker may even be a valid user.

Setup

# should install requirements.txt packages
$ pip install -e login-attempt-simulator # path to top level where setup.py is

# if not, install them explicitly
$ pip install -r requirements.txt

Usage

Create User Base and IPs

>>> import login_attempt_simulator as sim

>>> user_base_file = 'user_data/user_base.txt'
>>> user_ip_mapping_file = 'user_data/user_ips.json'

# create usernames and write to file
>>> sim.utils.make_user_base(user_base_file)

# create one or more IP addresses per user and save mapping to file
>>> valid_users = sim.utils.get_valid_users(user_base_file)
>>> sim.utils.save_user_ips(
...     sim.utils.assign_ip_addresses(valid_users),
...     user_ip_mapping_file
... )

Simulate Over Time Period

Note you can provide your own success probabilities when you instantiate the LoginAttemptSimulator.

>>> import datetime as dt

>>> start = dt.datetime(2019, 2, 1, 7, 0)
>>> end = start + dt.timedelta(days=10.5)

>>> simulator = sim.LoginAttemptSimulator(user_ip_mapping_file, start, end)
>>> simulator.simulate(attack_prob=0.01, try_all_users_prob=0.25, vary_ips=True)

Save Logs for Analysis Later

>>> simulator.save_hack_log('logs/attacks.csv')
>>> simulator.save_log('logs/log.csv')

About the Author

Stefanie Molin (@stefmolin) is a software engineer and data scientist at Bloomberg in New York City, where she tackles tough problems in information security, particularly those revolving around data wrangling/visualization, building tools for gathering data, and knowledge sharing. She is also the author of Hands-On Data Analysis with Pandas, which is currently in its second edition and has been translated into Korean. She holds a bachelor’s of science degree in operations research from Columbia University's Fu Foundation School of Engineering and Applied Science, as well as a master’s degree in computer science, with a specialization in machine learning, from Georgia Tech. In her free time, she enjoys traveling the world, inventing new recipes, and learning new languages spoken among both people and computers.

dvklopfenstein/login-attempt-simulator