| 1 |
//! Reed-Solomon Error Correction |
| 2 |
//! |
| 3 |
//! Efficient Reed-Solomon coding for ZephyrFS chunks, providing mathematical |
| 4 |
//! redundancy that can reconstruct data from partial chunks |
| 5 |
|
| 6 |
use anyhow::Result; |
| 7 |
use serde::{Deserialize, Serialize}; |
| 8 |
use std::collections::HashMap; |
| 9 |
use chrono::{DateTime, Utc}; |
| 10 |
|
| 11 |
/// Reed-Solomon encoder/decoder for ZephyrFS |
| 12 |
#[derive(Debug, Clone)] |
| 13 |
pub struct ReedSolomonCodec { |
| 14 |
/// Coding parameters |
| 15 |
pub config: ReedSolomonConfig, |
| 16 |
/// Galois field arithmetic |
| 17 |
pub galois_field: GaloisField, |
| 18 |
/// Encoding matrices |
| 19 |
pub encoding_matrix: Vec<Vec<u8>>, |
| 20 |
pub decoding_cache: HashMap<Vec<usize>, Vec<Vec<u8>>>, |
| 21 |
} |
| 22 |
|
| 23 |
#[derive(Debug, Clone, Serialize, Deserialize)] |
| 24 |
pub struct ReedSolomonConfig { |
| 25 |
/// Number of data chunks |
| 26 |
pub data_chunks: u32, |
| 27 |
/// Number of parity chunks |
| 28 |
pub parity_chunks: u32, |
| 29 |
/// Chunk size in bytes |
| 30 |
pub chunk_size: usize, |
| 31 |
/// Galois field size (typically 2^8 = 256) |
| 32 |
pub field_size: u32, |
| 33 |
/// Primitive polynomial for GF(2^8) |
| 34 |
pub primitive_poly: u32, |
| 35 |
} |
| 36 |
|
| 37 |
/// Galois Field arithmetic for Reed-Solomon |
| 38 |
#[derive(Debug, Clone)] |
| 39 |
pub struct GaloisField { |
| 40 |
/// Field size (2^8 = 256) |
| 41 |
pub size: u32, |
| 42 |
/// Generator polynomial |
| 43 |
pub generator: u32, |
| 44 |
/// Logarithm table |
| 45 |
pub log_table: Vec<u8>, |
| 46 |
/// Antilogarithm table |
| 47 |
pub antilog_table: Vec<u8>, |
| 48 |
} |
| 49 |
|
| 50 |
#[derive(Debug, Clone, Serialize, Deserialize)] |
| 51 |
pub struct EncodedChunk { |
| 52 |
pub chunk_id: String, |
| 53 |
pub chunk_index: u32, |
| 54 |
pub chunk_type: ChunkType, |
| 55 |
pub data: Vec<u8>, |
| 56 |
pub metadata: ChunkMetadata, |
| 57 |
pub created_at: DateTime<Utc>, |
| 58 |
} |
| 59 |
|
| 60 |
#[derive(Debug, Clone, Serialize, Deserialize)] |
| 61 |
pub enum ChunkType { |
| 62 |
Data, |
| 63 |
Parity, |
| 64 |
} |
| 65 |
|
| 66 |
#[derive(Debug, Clone, Serialize, Deserialize)] |
| 67 |
pub struct ChunkMetadata { |
| 68 |
pub original_size: usize, |
| 69 |
pub checksum: String, |
| 70 |
pub encoding_config: ReedSolomonConfig, |
| 71 |
pub chunk_position: u32, |
| 72 |
pub total_chunks: u32, |
| 73 |
} |
| 74 |
|
| 75 |
#[derive(Debug, Clone, Serialize, Deserialize)] |
| 76 |
pub struct ReconstructionRequest { |
| 77 |
pub original_chunk_id: String, |
| 78 |
pub available_chunks: Vec<EncodedChunk>, |
| 79 |
pub missing_chunk_indices: Vec<u32>, |
| 80 |
pub target_chunk_size: usize, |
| 81 |
} |
| 82 |
|
| 83 |
#[derive(Debug, Clone, Serialize, Deserialize)] |
| 84 |
pub struct ReconstructionResult { |
| 85 |
pub success: bool, |
| 86 |
pub reconstructed_chunks: Vec<EncodedChunk>, |
| 87 |
pub reconstruction_time_ms: u64, |
| 88 |
pub error_message: Option<String>, |
| 89 |
pub verification_passed: bool, |
| 90 |
} |
| 91 |
|
| 92 |
impl Default for ReedSolomonConfig { |
| 93 |
fn default() -> Self { |
| 94 |
Self { |
| 95 |
data_chunks: 6, |
| 96 |
parity_chunks: 3, |
| 97 |
chunk_size: 1024 * 1024, // 1MB chunks |
| 98 |
field_size: 256, |
| 99 |
primitive_poly: 0x11d, // Standard GF(2^8) primitive polynomial |
| 100 |
} |
| 101 |
} |
| 102 |
} |
| 103 |
|
| 104 |
impl GaloisField { |
| 105 |
/// Create new Galois Field GF(2^8) |
| 106 |
pub fn new() -> Self { |
| 107 |
let mut gf = Self { |
| 108 |
size: 256, |
| 109 |
generator: 0x11d, // x^8 + x^4 + x^3 + x^2 + 1 |
| 110 |
log_table: vec![0; 256], |
| 111 |
antilog_table: vec![0; 256], |
| 112 |
}; |
| 113 |
|
| 114 |
gf.build_tables(); |
| 115 |
gf |
| 116 |
} |
| 117 |
|
| 118 |
/// Build logarithm and antilogarithm tables |
| 119 |
fn build_tables(&mut self) { |
| 120 |
let mut val = 1u32; |
| 121 |
|
| 122 |
for i in 0..255 { |
| 123 |
self.antilog_table[i] = val as u8; |
| 124 |
self.log_table[val as usize] = i as u8; |
| 125 |
|
| 126 |
val <<= 1; |
| 127 |
if val & self.size != 0 { |
| 128 |
val ^= self.generator; |
| 129 |
} |
| 130 |
} |
| 131 |
|
| 132 |
self.antilog_table[255] = self.antilog_table[0]; |
| 133 |
} |
| 134 |
|
| 135 |
/// Galois field multiplication |
| 136 |
pub fn multiply(&self, a: u8, b: u8) -> u8 { |
| 137 |
if a == 0 || b == 0 { |
| 138 |
return 0; |
| 139 |
} |
| 140 |
|
| 141 |
let log_sum = (self.log_table[a as usize] as u32 + self.log_table[b as usize] as u32) % 255; |
| 142 |
self.antilog_table[log_sum as usize] |
| 143 |
} |
| 144 |
|
| 145 |
/// Galois field division |
| 146 |
pub fn divide(&self, a: u8, b: u8) -> u8 { |
| 147 |
if a == 0 { |
| 148 |
return 0; |
| 149 |
} |
| 150 |
if b == 0 { |
| 151 |
panic!("Division by zero in Galois field"); |
| 152 |
} |
| 153 |
|
| 154 |
let log_diff = (self.log_table[a as usize] as i32 - self.log_table[b as usize] as i32 + 255) % 255; |
| 155 |
self.antilog_table[log_diff as usize] |
| 156 |
} |
| 157 |
|
| 158 |
/// Galois field power |
| 159 |
pub fn power(&self, base: u8, exp: u8) -> u8 { |
| 160 |
if base == 0 { |
| 161 |
return 0; |
| 162 |
} |
| 163 |
|
| 164 |
let log_result = (self.log_table[base as usize] as u32 * exp as u32) % 255; |
| 165 |
self.antilog_table[log_result as usize] |
| 166 |
} |
| 167 |
} |
| 168 |
|
| 169 |
impl ReedSolomonCodec { |
| 170 |
/// Create new Reed-Solomon codec |
| 171 |
pub fn new(config: ReedSolomonConfig) -> Result<Self> { |
| 172 |
let galois_field = GaloisField::new(); |
| 173 |
let encoding_matrix = Self::build_encoding_matrix(&config, &galois_field)?; |
| 174 |
|
| 175 |
Ok(Self { |
| 176 |
config, |
| 177 |
galois_field, |
| 178 |
encoding_matrix, |
| 179 |
decoding_cache: HashMap::new(), |
| 180 |
}) |
| 181 |
} |
| 182 |
|
| 183 |
/// Build encoding matrix for Reed-Solomon |
| 184 |
fn build_encoding_matrix(config: &ReedSolomonConfig, gf: &GaloisField) -> Result<Vec<Vec<u8>>> { |
| 185 |
let total_chunks = config.data_chunks + config.parity_chunks; |
| 186 |
let mut matrix = Vec::with_capacity(total_chunks as usize); |
| 187 |
|
| 188 |
// Identity matrix for data chunks |
| 189 |
for i in 0..config.data_chunks { |
| 190 |
let mut row = vec![0u8; config.data_chunks as usize]; |
| 191 |
row[i as usize] = 1; |
| 192 |
matrix.push(row); |
| 193 |
} |
| 194 |
|
| 195 |
// Vandermonde matrix for parity chunks |
| 196 |
for i in 0..config.parity_chunks { |
| 197 |
let mut row = Vec::with_capacity(config.data_chunks as usize); |
| 198 |
let alpha = (i + 1) as u8; // Generator element |
| 199 |
|
| 200 |
for j in 0..config.data_chunks { |
| 201 |
row.push(gf.power(alpha, j as u8)); |
| 202 |
} |
| 203 |
matrix.push(row); |
| 204 |
} |
| 205 |
|
| 206 |
Ok(matrix) |
| 207 |
} |
| 208 |
|
| 209 |
/// Encode data into Reed-Solomon chunks |
| 210 |
pub fn encode(&self, chunk_id: String, data: &[u8]) -> Result<Vec<EncodedChunk>> { |
| 211 |
let chunk_size = self.config.chunk_size; |
| 212 |
let data_chunks = self.config.data_chunks as usize; |
| 213 |
let total_chunks = (self.config.data_chunks + self.config.parity_chunks) as usize; |
| 214 |
|
| 215 |
// Pad data to fit evenly into data chunks |
| 216 |
let padded_size = ((data.len() + data_chunks - 1) / data_chunks) * data_chunks; |
| 217 |
let mut padded_data = data.to_vec(); |
| 218 |
padded_data.resize(padded_size, 0); |
| 219 |
|
| 220 |
let chunk_data_size = padded_size / data_chunks; |
| 221 |
let mut chunks = Vec::with_capacity(total_chunks); |
| 222 |
|
| 223 |
// Create data chunks |
| 224 |
for i in 0..data_chunks { |
| 225 |
let start = i * chunk_data_size; |
| 226 |
let end = start + chunk_data_size; |
| 227 |
let chunk_data = padded_data[start..end].to_vec(); |
| 228 |
|
| 229 |
chunks.push(EncodedChunk { |
| 230 |
chunk_id: format!("{}_{}", chunk_id, i), |
| 231 |
chunk_index: i as u32, |
| 232 |
chunk_type: ChunkType::Data, |
| 233 |
data: chunk_data, |
| 234 |
metadata: ChunkMetadata { |
| 235 |
original_size: if i == data_chunks - 1 { |
| 236 |
data.len() - (data_chunks - 1) * chunk_data_size |
| 237 |
} else { |
| 238 |
chunk_data_size |
| 239 |
}, |
| 240 |
checksum: self.calculate_checksum(&padded_data[start..end]), |
| 241 |
encoding_config: self.config.clone(), |
| 242 |
chunk_position: i as u32, |
| 243 |
total_chunks: total_chunks as u32, |
| 244 |
}, |
| 245 |
created_at: Utc::now(), |
| 246 |
}); |
| 247 |
} |
| 248 |
|
| 249 |
// Create parity chunks |
| 250 |
for i in data_chunks..total_chunks { |
| 251 |
let mut parity_data = vec![0u8; chunk_data_size]; |
| 252 |
|
| 253 |
// Calculate parity using encoding matrix |
| 254 |
for j in 0..chunk_data_size { |
| 255 |
let mut parity_byte = 0u8; |
| 256 |
|
| 257 |
for k in 0..data_chunks { |
| 258 |
let data_byte = padded_data[k * chunk_data_size + j]; |
| 259 |
let coeff = self.encoding_matrix[i][k]; |
| 260 |
parity_byte ^= self.galois_field.multiply(data_byte, coeff); |
| 261 |
} |
| 262 |
|
| 263 |
parity_data[j] = parity_byte; |
| 264 |
} |
| 265 |
|
| 266 |
chunks.push(EncodedChunk { |
| 267 |
chunk_id: format!("{}_{}", chunk_id, i), |
| 268 |
chunk_index: i as u32, |
| 269 |
chunk_type: ChunkType::Parity, |
| 270 |
data: parity_data.clone(), |
| 271 |
metadata: ChunkMetadata { |
| 272 |
original_size: chunk_data_size, |
| 273 |
checksum: self.calculate_checksum(&parity_data), |
| 274 |
encoding_config: self.config.clone(), |
| 275 |
chunk_position: i as u32, |
| 276 |
total_chunks: total_chunks as u32, |
| 277 |
}, |
| 278 |
created_at: Utc::now(), |
| 279 |
}); |
| 280 |
} |
| 281 |
|
| 282 |
Ok(chunks) |
| 283 |
} |
| 284 |
|
| 285 |
/// Decode/reconstruct data from available chunks |
| 286 |
pub fn decode(&mut self, request: ReconstructionRequest) -> Result<ReconstructionResult> { |
| 287 |
let start_time = crate::SerializableInstant::now(); |
| 288 |
|
| 289 |
// Verify we have enough chunks for reconstruction |
| 290 |
if request.available_chunks.len() < self.config.data_chunks as usize { |
| 291 |
return Ok(ReconstructionResult { |
| 292 |
success: false, |
| 293 |
reconstructed_chunks: Vec::new(), |
| 294 |
reconstruction_time_ms: start_time.elapsed().as_millis() as u64, |
| 295 |
error_message: Some(format!( |
| 296 |
"Insufficient chunks: need {}, have {}", |
| 297 |
self.config.data_chunks, |
| 298 |
request.available_chunks.len() |
| 299 |
)), |
| 300 |
verification_passed: false, |
| 301 |
}); |
| 302 |
} |
| 303 |
|
| 304 |
let data_chunks = self.config.data_chunks as usize; |
| 305 |
let chunk_size = request.available_chunks[0].data.len(); |
| 306 |
|
| 307 |
// Build decoding matrix |
| 308 |
let available_indices: Vec<usize> = request.available_chunks |
| 309 |
.iter() |
| 310 |
.map(|chunk| chunk.chunk_index as usize) |
| 311 |
.collect(); |
| 312 |
|
| 313 |
let decoding_matrix = self.get_or_build_decoding_matrix(&available_indices)?; |
| 314 |
|
| 315 |
// Reconstruct missing data |
| 316 |
let mut reconstructed_data = vec![vec![0u8; chunk_size]; data_chunks]; |
| 317 |
|
| 318 |
// Copy available data chunks |
| 319 |
for chunk in &request.available_chunks { |
| 320 |
let idx = chunk.chunk_index as usize; |
| 321 |
if idx < data_chunks { |
| 322 |
reconstructed_data[idx] = chunk.data.clone(); |
| 323 |
} |
| 324 |
} |
| 325 |
|
| 326 |
// Reconstruct missing data chunks |
| 327 |
for missing_idx in &request.missing_chunk_indices { |
| 328 |
let missing_idx = *missing_idx as usize; |
| 329 |
if missing_idx >= data_chunks { |
| 330 |
continue; // Skip parity chunks for now |
| 331 |
} |
| 332 |
|
| 333 |
let mut reconstructed_chunk = vec![0u8; chunk_size]; |
| 334 |
|
| 335 |
for byte_pos in 0..chunk_size { |
| 336 |
let mut reconstructed_byte = 0u8; |
| 337 |
|
| 338 |
for (matrix_col, chunk) in request.available_chunks.iter().enumerate().take(data_chunks) { |
| 339 |
let coeff = decoding_matrix[missing_idx][matrix_col]; |
| 340 |
let data_byte = chunk.data[byte_pos]; |
| 341 |
reconstructed_byte ^= self.galois_field.multiply(data_byte, coeff); |
| 342 |
} |
| 343 |
|
| 344 |
reconstructed_chunk[byte_pos] = reconstructed_byte; |
| 345 |
} |
| 346 |
|
| 347 |
reconstructed_data[missing_idx] = reconstructed_chunk; |
| 348 |
} |
| 349 |
|
| 350 |
// Create reconstructed chunk objects |
| 351 |
let mut reconstructed_chunks = Vec::new(); |
| 352 |
for missing_idx in &request.missing_chunk_indices { |
| 353 |
let idx = *missing_idx as usize; |
| 354 |
if idx < data_chunks { |
| 355 |
reconstructed_chunks.push(EncodedChunk { |
| 356 |
chunk_id: format!("{}_{}", request.original_chunk_id, idx), |
| 357 |
chunk_index: idx as u32, |
| 358 |
chunk_type: ChunkType::Data, |
| 359 |
data: reconstructed_data[idx].clone(), |
| 360 |
metadata: ChunkMetadata { |
| 361 |
original_size: chunk_size, |
| 362 |
checksum: self.calculate_checksum(&reconstructed_data[idx]), |
| 363 |
encoding_config: self.config.clone(), |
| 364 |
chunk_position: idx as u32, |
| 365 |
total_chunks: (self.config.data_chunks + self.config.parity_chunks), |
| 366 |
}, |
| 367 |
created_at: Utc::now(), |
| 368 |
}); |
| 369 |
} |
| 370 |
} |
| 371 |
|
| 372 |
// Verify reconstruction |
| 373 |
let verification_passed = self.verify_reconstruction(&reconstructed_chunks)?; |
| 374 |
|
| 375 |
Ok(ReconstructionResult { |
| 376 |
success: true, |
| 377 |
reconstructed_chunks, |
| 378 |
reconstruction_time_ms: start_time.elapsed().as_millis() as u64, |
| 379 |
error_message: None, |
| 380 |
verification_passed, |
| 381 |
}) |
| 382 |
} |
| 383 |
|
| 384 |
/// Get or build decoding matrix for given available chunks |
| 385 |
fn get_or_build_decoding_matrix(&mut self, available_indices: &[usize]) -> Result<Vec<Vec<u8>>> { |
| 386 |
// Check cache first |
| 387 |
if let Some(cached_matrix) = self.decoding_cache.get(available_indices) { |
| 388 |
return Ok(cached_matrix.clone()); |
| 389 |
} |
| 390 |
|
| 391 |
let data_chunks = self.config.data_chunks as usize; |
| 392 |
|
| 393 |
// Build submatrix from encoding matrix using available chunks |
| 394 |
let mut submatrix = Vec::with_capacity(data_chunks); |
| 395 |
for &idx in available_indices.iter().take(data_chunks) { |
| 396 |
submatrix.push(self.encoding_matrix[idx].clone()); |
| 397 |
} |
| 398 |
|
| 399 |
// Invert the submatrix |
| 400 |
let decoding_matrix = self.invert_matrix(submatrix)?; |
| 401 |
|
| 402 |
// Cache the result |
| 403 |
self.decoding_cache.insert(available_indices.to_vec(), decoding_matrix.clone()); |
| 404 |
|
| 405 |
Ok(decoding_matrix) |
| 406 |
} |
| 407 |
|
| 408 |
/// Invert a matrix in Galois field |
| 409 |
fn invert_matrix(&self, mut matrix: Vec<Vec<u8>>) -> Result<Vec<Vec<u8>>> { |
| 410 |
let n = matrix.len(); |
| 411 |
|
| 412 |
// Create augmented matrix [A|I] |
| 413 |
let mut augmented = Vec::with_capacity(n); |
| 414 |
for i in 0..n { |
| 415 |
let mut row = matrix[i].clone(); |
| 416 |
for j in 0..n { |
| 417 |
row.push(if i == j { 1 } else { 0 }); |
| 418 |
} |
| 419 |
augmented.push(row); |
| 420 |
} |
| 421 |
|
| 422 |
// Gaussian elimination |
| 423 |
for i in 0..n { |
| 424 |
// Find pivot |
| 425 |
let mut pivot_row = i; |
| 426 |
for j in (i + 1)..n { |
| 427 |
if augmented[j][i] != 0 { |
| 428 |
pivot_row = j; |
| 429 |
break; |
| 430 |
} |
| 431 |
} |
| 432 |
|
| 433 |
if augmented[pivot_row][i] == 0 { |
| 434 |
return Err(anyhow::anyhow!("Matrix is not invertible")); |
| 435 |
} |
| 436 |
|
| 437 |
// Swap rows if needed |
| 438 |
if pivot_row != i { |
| 439 |
augmented.swap(i, pivot_row); |
| 440 |
} |
| 441 |
|
| 442 |
// Scale pivot row |
| 443 |
let pivot = augmented[i][i]; |
| 444 |
for j in 0..(2 * n) { |
| 445 |
augmented[i][j] = self.galois_field.divide(augmented[i][j], pivot); |
| 446 |
} |
| 447 |
|
| 448 |
// Eliminate column |
| 449 |
for j in 0..n { |
| 450 |
if i != j && augmented[j][i] != 0 { |
| 451 |
let factor = augmented[j][i]; |
| 452 |
for k in 0..(2 * n) { |
| 453 |
let product = self.galois_field.multiply(factor, augmented[i][k]); |
| 454 |
augmented[j][k] ^= product; |
| 455 |
} |
| 456 |
} |
| 457 |
} |
| 458 |
} |
| 459 |
|
| 460 |
// Extract inverse matrix |
| 461 |
let mut inverse = Vec::with_capacity(n); |
| 462 |
for i in 0..n { |
| 463 |
inverse.push(augmented[i][n..].to_vec()); |
| 464 |
} |
| 465 |
|
| 466 |
Ok(inverse) |
| 467 |
} |
| 468 |
|
| 469 |
/// Calculate checksum for data |
| 470 |
fn calculate_checksum(&self, data: &[u8]) -> String { |
| 471 |
// Simple checksum - in production would use cryptographic hash |
| 472 |
let sum: u32 = data.iter().map(|&b| b as u32).sum(); |
| 473 |
format!("{:08x}", sum) |
| 474 |
} |
| 475 |
|
| 476 |
/// Verify reconstruction correctness |
| 477 |
fn verify_reconstruction(&self, chunks: &[EncodedChunk]) -> Result<bool> { |
| 478 |
// Verify checksums |
| 479 |
for chunk in chunks { |
| 480 |
let calculated_checksum = self.calculate_checksum(&chunk.data); |
| 481 |
if calculated_checksum != chunk.metadata.checksum { |
| 482 |
return Ok(false); |
| 483 |
} |
| 484 |
} |
| 485 |
|
| 486 |
// Additional verification could include re-encoding and comparing |
| 487 |
Ok(true) |
| 488 |
} |
| 489 |
|
| 490 |
/// Get redundancy schemes available |
| 491 |
pub fn get_available_schemes() -> Vec<ReedSolomonConfig> { |
| 492 |
vec![ |
| 493 |
// Standard schemes |
| 494 |
ReedSolomonConfig { |
| 495 |
data_chunks: 3, |
| 496 |
parity_chunks: 2, |
| 497 |
chunk_size: 1024 * 1024, |
| 498 |
field_size: 256, |
| 499 |
primitive_poly: 0x11d, |
| 500 |
}, |
| 501 |
ReedSolomonConfig { |
| 502 |
data_chunks: 6, |
| 503 |
parity_chunks: 3, |
| 504 |
chunk_size: 1024 * 1024, |
| 505 |
field_size: 256, |
| 506 |
primitive_poly: 0x11d, |
| 507 |
}, |
| 508 |
ReedSolomonConfig { |
| 509 |
data_chunks: 10, |
| 510 |
parity_chunks: 4, |
| 511 |
chunk_size: 1024 * 1024, |
| 512 |
field_size: 256, |
| 513 |
primitive_poly: 0x11d, |
| 514 |
}, |
| 515 |
// High redundancy for critical data |
| 516 |
ReedSolomonConfig { |
| 517 |
data_chunks: 8, |
| 518 |
parity_chunks: 6, |
| 519 |
chunk_size: 1024 * 1024, |
| 520 |
field_size: 256, |
| 521 |
primitive_poly: 0x11d, |
| 522 |
}, |
| 523 |
] |
| 524 |
} |
| 525 |
|
| 526 |
/// Calculate storage efficiency for a scheme |
| 527 |
pub fn calculate_storage_efficiency(&self) -> f64 { |
| 528 |
self.config.data_chunks as f64 / (self.config.data_chunks + self.config.parity_chunks) as f64 |
| 529 |
} |
| 530 |
|
| 531 |
/// Calculate fault tolerance |
| 532 |
pub fn calculate_fault_tolerance(&self) -> u32 { |
| 533 |
self.config.parity_chunks |
| 534 |
} |
| 535 |
|
| 536 |
/// Estimate reconstruction performance |
| 537 |
pub fn estimate_reconstruction_time(&self, chunk_size_mb: f64, available_bandwidth_mbps: f64) -> f64 { |
| 538 |
let data_to_transfer = chunk_size_mb * self.config.data_chunks as f64; |
| 539 |
let transfer_time = data_to_transfer / available_bandwidth_mbps; |
| 540 |
let computation_time = chunk_size_mb * 0.1; // Estimate 0.1 seconds per MB for computation |
| 541 |
|
| 542 |
transfer_time + computation_time |
| 543 |
} |
| 544 |
} |
| 545 |
|
| 546 |
/// Reed-Solomon manager for handling multiple coding schemes |
| 547 |
#[derive(Debug)] |
| 548 |
pub struct ReedSolomonManager { |
| 549 |
pub codecs: HashMap<String, ReedSolomonCodec>, |
| 550 |
pub default_scheme: String, |
| 551 |
} |
| 552 |
|
| 553 |
impl ReedSolomonManager { |
| 554 |
/// Create new Reed-Solomon manager with multiple schemes |
| 555 |
pub fn new() -> Result<Self> { |
| 556 |
let mut manager = Self { |
| 557 |
codecs: HashMap::new(), |
| 558 |
default_scheme: "6+3".to_string(), |
| 559 |
}; |
| 560 |
|
| 561 |
// Initialize standard schemes |
| 562 |
let schemes = [ |
| 563 |
("3+2", ReedSolomonConfig { data_chunks: 3, parity_chunks: 2, ..Default::default() }), |
| 564 |
("6+3", ReedSolomonConfig { data_chunks: 6, parity_chunks: 3, ..Default::default() }), |
| 565 |
("10+4", ReedSolomonConfig { data_chunks: 10, parity_chunks: 4, ..Default::default() }), |
| 566 |
("8+6", ReedSolomonConfig { data_chunks: 8, parity_chunks: 6, ..Default::default() }), |
| 567 |
]; |
| 568 |
|
| 569 |
for (name, config) in schemes { |
| 570 |
let codec = ReedSolomonCodec::new(config)?; |
| 571 |
manager.codecs.insert(name.to_string(), codec); |
| 572 |
} |
| 573 |
|
| 574 |
Ok(manager) |
| 575 |
} |
| 576 |
|
| 577 |
/// Get codec for scheme |
| 578 |
pub fn get_codec(&mut self, scheme: &str) -> Result<&mut ReedSolomonCodec> { |
| 579 |
self.codecs.get_mut(scheme) |
| 580 |
.ok_or_else(|| anyhow::anyhow!("Reed-Solomon scheme '{}' not found", scheme)) |
| 581 |
} |
| 582 |
|
| 583 |
/// Recommend scheme based on requirements |
| 584 |
pub fn recommend_scheme(&self, durability_requirement: f64, cost_sensitivity: f64) -> String { |
| 585 |
if durability_requirement > 0.99999 { |
| 586 |
"8+6".to_string() // Ultra high durability |
| 587 |
} else if durability_requirement > 0.9999 { |
| 588 |
"10+4".to_string() // High durability |
| 589 |
} else if cost_sensitivity > 0.7 { |
| 590 |
"3+2".to_string() // Cost sensitive |
| 591 |
} else { |
| 592 |
"6+3".to_string() // Balanced |
| 593 |
} |
| 594 |
} |
| 595 |
|
| 596 |
/// Get scheme performance characteristics |
| 597 |
pub fn get_scheme_info(&self, scheme: &str) -> Option<SchemeInfo> { |
| 598 |
self.codecs.get(scheme).map(|codec| { |
| 599 |
SchemeInfo { |
| 600 |
name: scheme.to_string(), |
| 601 |
data_chunks: codec.config.data_chunks, |
| 602 |
parity_chunks: codec.config.parity_chunks, |
| 603 |
storage_efficiency: codec.calculate_storage_efficiency(), |
| 604 |
fault_tolerance: codec.calculate_fault_tolerance(), |
| 605 |
reconstruction_complexity: (codec.config.data_chunks * codec.config.parity_chunks) as f64, |
| 606 |
} |
| 607 |
}) |
| 608 |
} |
| 609 |
} |
| 610 |
|
| 611 |
#[derive(Debug, Clone, Serialize, Deserialize)] |
| 612 |
pub struct SchemeInfo { |
| 613 |
pub name: String, |
| 614 |
pub data_chunks: u32, |
| 615 |
pub parity_chunks: u32, |
| 616 |
pub storage_efficiency: f64, |
| 617 |
pub fault_tolerance: u32, |
| 618 |
pub reconstruction_complexity: f64, |
| 619 |
} |
| 620 |
|
| 621 |
#[cfg(test)] |
| 622 |
mod tests { |
| 623 |
use super::*; |
| 624 |
|
| 625 |
#[test] |
| 626 |
fn test_galois_field_arithmetic() { |
| 627 |
let gf = GaloisField::new(); |
| 628 |
|
| 629 |
// Test basic properties |
| 630 |
assert_eq!(gf.multiply(0, 5), 0); |
| 631 |
assert_eq!(gf.multiply(1, 5), 5); |
| 632 |
assert_eq!(gf.divide(10, 2), gf.multiply(10, gf.divide(1, 2))); |
| 633 |
|
| 634 |
// Test multiplicative inverse |
| 635 |
for i in 1..=255u8 { |
| 636 |
let inverse = gf.divide(1, i); |
| 637 |
assert_eq!(gf.multiply(i, inverse), 1); |
| 638 |
} |
| 639 |
} |
| 640 |
|
| 641 |
#[test] |
| 642 |
fn test_reed_solomon_codec_creation() { |
| 643 |
let config = ReedSolomonConfig::default(); |
| 644 |
let codec = ReedSolomonCodec::new(config).unwrap(); |
| 645 |
|
| 646 |
assert_eq!(codec.config.data_chunks, 6); |
| 647 |
assert_eq!(codec.config.parity_chunks, 3); |
| 648 |
assert_eq!(codec.encoding_matrix.len(), 9); |
| 649 |
} |
| 650 |
|
| 651 |
#[test] |
| 652 |
fn test_encoding_and_reconstruction() { |
| 653 |
let config = ReedSolomonConfig { |
| 654 |
data_chunks: 3, |
| 655 |
parity_chunks: 2, |
| 656 |
chunk_size: 1024, |
| 657 |
..Default::default() |
| 658 |
}; |
| 659 |
|
| 660 |
let mut codec = ReedSolomonCodec::new(config).unwrap(); |
| 661 |
|
| 662 |
// Test data |
| 663 |
let test_data = b"Hello, Reed-Solomon World! This is a test of error correction.".to_vec(); |
| 664 |
|
| 665 |
// Encode |
| 666 |
let chunks = codec.encode("test_chunk".to_string(), &test_data).unwrap(); |
| 667 |
assert_eq!(chunks.len(), 5); // 3 data + 2 parity |
| 668 |
|
| 669 |
// Simulate losing 2 chunks (within error correction capability) |
| 670 |
let available_chunks = vec![chunks[0].clone(), chunks[2].clone(), chunks[4].clone()]; |
| 671 |
let missing_indices = vec![1, 3]; |
| 672 |
|
| 673 |
// Reconstruct |
| 674 |
let request = ReconstructionRequest { |
| 675 |
original_chunk_id: "test_chunk".to_string(), |
| 676 |
available_chunks, |
| 677 |
missing_chunk_indices: missing_indices, |
| 678 |
target_chunk_size: 1024, |
| 679 |
}; |
| 680 |
|
| 681 |
let result = codec.decode(request).unwrap(); |
| 682 |
assert!(result.success); |
| 683 |
assert_eq!(result.reconstructed_chunks.len(), 2); |
| 684 |
assert!(result.verification_passed); |
| 685 |
} |
| 686 |
|
| 687 |
#[test] |
| 688 |
fn test_manager_scheme_recommendation() { |
| 689 |
let manager = ReedSolomonManager::new().unwrap(); |
| 690 |
|
| 691 |
// High durability requirement |
| 692 |
assert_eq!(manager.recommend_scheme(0.99999, 0.3), "8+6"); |
| 693 |
|
| 694 |
// Cost sensitive |
| 695 |
assert_eq!(manager.recommend_scheme(0.999, 0.8), "3+2"); |
| 696 |
|
| 697 |
// Balanced |
| 698 |
assert_eq!(manager.recommend_scheme(0.999, 0.3), "6+3"); |
| 699 |
} |
| 700 |
|
| 701 |
#[test] |
| 702 |
fn test_storage_efficiency_calculation() { |
| 703 |
let config = ReedSolomonConfig { |
| 704 |
data_chunks: 6, |
| 705 |
parity_chunks: 3, |
| 706 |
..Default::default() |
| 707 |
}; |
| 708 |
|
| 709 |
let codec = ReedSolomonCodec::new(config).unwrap(); |
| 710 |
let efficiency = codec.calculate_storage_efficiency(); |
| 711 |
|
| 712 |
assert!((efficiency - 0.6667).abs() < 0.01); // 6/9 ≈ 0.6667 |
| 713 |
} |
| 714 |
} |