Registrations for the July 2021 System Design Cohort are now open Enroll Today

Fork Bomb

Published on 9th Jun 2021

2 min read

Share this article on


In this essay, we explore a simple yet effective DoS attack called Fork Bomb, also called Rabbit Virus. This attack forks out processes infinitely, starving them for any resources.

Online Coding Platforms and Code Evaluators are susceptible to this attack as they accept raw code from the user and execute it. So, if you are building one, do ensure you are protected against it and infinite loops. We will also discuss how to prevent and defend against Fork Bomb in the final section of this essay.

How do they work?

Fork bombs can be summarized in just three words: Fork until possible. When executed, fork bombs continuously fork out non-terminating processes, demanding machine resources like CPU (mostly) and memory.

Fork Bomb

With so many processes competing for the CPU and other resources (if provisioned), the scheduler and CPU are put under tremendous load. After a specific limit, the entire system stalls.

Implementing Fork Bombs

Before we take a look at how to prevent or stop a Fork Bomb, let's look at something more interesting - how to implement a Fork Bomb?

A quick detour, let's see what fork does: Upon every invocation, the forked child process is an exact duplicate of the parent process except for a few details, but nonetheless what matters to us is that it runs the exact same code as the parent.

C implementation

A simple C implementation of a Fork Bomb could be, to fork child processes within an infinite for loop, resulting in exponential forking of child processes.

#include <unistd.h>
int main(void) {
    for (;;) {
        fork();
    }
}

With the fork being invoked inside the infinite for loop, every single child process and the parent process will continue to remain stuck in the infinite loop while continuously forking out more and more child processes that execute the same code and stuck in the same loop; and thus resulting in exponential child forks.

These child processes start consuming the resources and blocking the legitimate programs. This prevents the creation of any new processes. This also freezes the process that responds to Keystrokes, putting the entire system to a standstill.

Bash implementation

There is a very famous Fork Bomb implementation in Bash, and the code that does this has no alphabets or numbers in it, just pure symbols.

:(){ :|:& };:

Although the shell statement looks gibberish, it is effortless to understand. In the statement above, we are defining a function named : having body :|:& and at the end invoking it using the name :, just like any usual shell function.

As part of the function body, we are again invoking the same function : and piping its output to the input of a background process executing another instance of :. This way, we are recursively invoking the same function (command) and stalling it by creating a pipe between the two.

A cleaner way to redefine this very implementation of Fork Bomb would be

bomb() { 
    bomb| bomb& 
};bomb

How to prevent them?

To protect our system against Fork Bombs, we can cap the processes owned by a certain user, thus blocking process creation at that cap.

Using the *nix utility called ulimit, we can set the maximum number of processes that a user can execute in the system, using the flag -u. By setting this value to an appropriate (lower) value, we can cap the process creation for a user, ensuring we can never be fork bombed by that user.

References


If my work adds value, consider supporting me


Buy Me A Coffee

Arpit's Newsletter

1300+ Signups

If you like what you read subscribe you can always subscribe to my newsletter and get the post delivered straight to your inbox. I write essays on various engineering topics and share it through my weekly newsletter 👇



Other articles that you might like

Publish python package on PyPI

Publish python package on PyPI

If you have written something cool in Python and want to make it installable via pip and easy_instal...

10th Nov
Better Ranking using Bayesian Average

Better Ranking using Bayesian Average

Ranking a list of movies, products, books or even restaurants is tricky and in this article, we find...

12th Apr
Fast and Efficient Pagination in MongoDB

Fast and Efficient Pagination in MongoDB

MongoDB is a document based data store and hence pagination is one of the most common use case of it...

6th Jun
Personalize your Python Prompt

Personalize your Python Prompt

Personalization is what we all love. In this article we find how we could personalize the Python int...

21st Feb


  • Made by Arpit Bhayani © 2021