Raptor codes are systematic fountain codes that achieve near-optimal performance by combining a high-rate pre-code (typically LDPC) with LT codes. They provide excellent error correction with minimal overhead and linear encoding/decoding complexity, making them suitable for broadcast and multicast applications.
| Property | Value |
|---|---|
| Category | Error Correction |
| Sub-category | Fountain Codes |
| Security status | 🛡️ Secure |
| Complexity | Expert |
| Inventor | Amin Shokrollahi |
| Year | 2006 |
| Origin | 🇨🇭 Switzerland |
| Source | algorithms/ecc/raptor-codes.js |
| Parameter | Supported values |
|---|---|
| Block sizes | 4 bytes (32 bits) to 1048576 bytes (8388608 bits) |
| Flag | Value |
|---|---|
supportsContinuousEncoding |
Yes |
supportsRateless |
Yes |
isSystematic |
Yes |
Status: 🛡️ Secure
No vulnerabilities are recorded for this implementation.
2 vectors ship with this algorithm and run in the test suite. Byte values are hexadecimal.
Vector 1 — Raptor encoding test with 8 source symbols
Source: Reference implementation test vector
| Field | Value |
|---|---|
k |
8 |
preCodeRate |
0.95 |
targetOverhead |
0.05 |
seed |
42 |
input |
48656c6c6f20576f |
expected |
48656c6c6f20576f03 |
Vector 2 — Raptor encoding test with 64 source symbols
Source: Reference implementation test vector
| Field | Value |
|---|---|
k |
64 |
preCodeRate |
0.9 |
targetOverhead |
0.1 |
seed |
314159 |
input |
000102030405060708090a0b0c0d0e0f 101112131415161718191a1b1c1d1e1f 202122232425262728292a2b2c2d2e2f 303132333435363738393a3b3c3d3e3f |
expected |
000102030405060708090a0b0c0d0e0f 101112131415161718191a1b1c1d1e1f 202122232425262728292a2b2c2d2e2f 303132333435363738393a3b3c3d3e3f 333e1e191e110701 |