odin-erasure

Library for erasure coding using binary finite fields. The code was originally based on vishesh-khemani/erasure-coding.

To test it out on the command line:

1. Install task from taskfile.dev
2. task run erasure encode [<options>] --file <input file> --code <code file prefix>
3. task run erasure decode [<options>] --file <output file> --code <code file prefix>

The options are as follows:

  -N | --num-code     number of code chunks in a block; default 5
  -K | --num-data     number of data chunks in a block; default 3
  -w | --word-size    number of bytes in each word in a chunks (1|2|4|8); default: 8

Note that you have to make sure to use the same <code file prefix> for encode and decode operations.

The default options will specify a 3 of 5 erasure coder (N=5, K=3). In other words, it will spread the data into 5 chunks but only requires any 3 out of 5 chunks to reconstruct the data.