https://github.com/hjfitz/disparp-mtp-cracker
"one-time pad on the back of your pinto is redneck" - Bruce Potter, 2012
https://github.com/hjfitz/disparp-mtp-cracker
dictionary-attack distributed one-time-pad parallel ts-node typescript worker-threads
Last synced: 4 months ago
JSON representation
"one-time pad on the back of your pinto is redneck" - Bruce Potter, 2012
- Host: GitHub
- URL: https://github.com/hjfitz/disparp-mtp-cracker
- Owner: hjfitz
- Created: 2019-04-13T13:40:42.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2019-04-26T10:32:32.000Z (about 6 years ago)
- Last Synced: 2025-01-14T03:38:51.497Z (5 months ago)
- Topics: dictionary-attack, distributed, one-time-pad, parallel, ts-node, typescript, worker-threads
- Language: TypeScript
- Size: 1.2 MB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# [DISPARP] Distributed Systems and Parallel Programming - Coursework Project
> University of Portsmouth, Year 3 Coursework
## Many-time Pad Reuse brute-forcing
This repo has code for both a distributed and parallel (via worker_threads) for breaking ciphertexts enciphered using OTP and a repeated dictionary work as a key.
## Requirements
Only Node and NPM are required to use this program. The dictionary is given (although you can replace this!)
* Node: >= 11
* NPM: >= 6.4.1
* Yarn (preferred over NPM): >= 1.12.3
## Usage
0. Install all dependencies `npm install`
1. Build the TypeScript source code to JS: `npm run build`
2. Run the code:
* For the distributed version, stand up the server with `npm run start:cluster-naster 2`. The number of clients required is the first argument - defaulting to 2.
* Stand up clients with `npm run start:cluster-client-client 192.168.0.12:5000`. The only argument defines where the server is hosted.
* For the parallel version, run `npm run start:threaded 2`. The only argument is the number of threads to spawn.
* To run the worker farm code, prepend the aforementioned commands with `farm-`, `npm run start:farm-cluster-worker` etc.
## Tech
Both versions run the same way. Clients are invoked with arguments:
* nodes in the pool
* node number
* ciphertext
Nodes then get their words from a local dictionary and split this by their calculated block size, and their position (Block size = wordlist length / number of nodes).
Nodes iterate through this, calculating (ciphertext ^ word). If the result is readable (defined by some regex: `/^[a-zA-Z0-9\., \'\"\-_\:\(\)]+$/`) then the key and decoded, readable, ciphertext is sent back to the host. If these two the CT and Key defined by the host, the program exits.
### Distributed
A websocket server waits for the desired amount of clients. Upon reaching this number, a 'begin' event is sent to each client, as well as their number, size and ciphertext (defined by the interface in lib/shared.ts). Nodes do their brute-force loop and send back any candidates.
### Threaded
Using `worker_threads`, threads are spawned according to the argument given to `threaded.js`. Worker threads are spawned with the `new Worker(workerfile, workerData)` constructor. Amount of threads, thread position and ciphertext is sent in workerData.
## Results
Available on [Google Sheets](https://docs.google.com/spreadsheets/d/1rfd0MSvlGD4TiAJzQ9l7tq3FgKl0P2Nm5BpdcP8GT0Q/edit#gid=1020168994)