Ashlesh Gawande | 0b2897e | 2018-06-20 14:40:47 -0500 | [diff] [blame] | 1 | /* -*- Mode:C++; c-file-style:"gnu"; indent-tabs-mode:nil; -*- */ |
| 2 | /* |
| 3 | * Copyright (c) 2014-2018, The University of Memphis |
| 4 | * |
| 5 | * This file is part of PSync. |
| 6 | * See AUTHORS.md for complete list of PSync authors and contributors. |
| 7 | * |
| 8 | * PSync is free software: you can redistribute it and/or modify it under the terms |
| 9 | * of the GNU General Public License as published by the Free Software Foundation, |
| 10 | * either version 3 of the License, or (at your option) any later version. |
| 11 | * |
| 12 | * PSync is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; |
| 13 | * without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR |
| 14 | * PURPOSE. See the GNU General Public License for more details. |
| 15 | * |
| 16 | * You should have received a copy of the GNU General Public License along with |
| 17 | * PSync, e.g., in COPYING.md file. If not, see <http://www.gnu.org/licenses/>. |
| 18 | * |
| 19 | |
| 20 | * This file incorporates work covered by the following copyright and |
| 21 | * permission notice: |
| 22 | |
| 23 | * The MIT License (MIT) |
| 24 | |
| 25 | * Copyright (c) 2014 Gavin Andresen |
| 26 | |
| 27 | * Permission is hereby granted, free of charge, to any person obtaining a copy |
| 28 | * of this software and associated documentation files (the "Software"), to deal |
| 29 | * in the Software without restriction, including without limitation the rights |
| 30 | * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell |
| 31 | * copies of the Software, and to permit persons to whom the Software is |
| 32 | * furnished to do so, subject to the following conditions: |
| 33 | |
| 34 | * The above copyright notice and this permission notice shall be included in all |
| 35 | * copies or substantial portions of the Software. |
| 36 | |
| 37 | * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR |
| 38 | * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
| 39 | * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE |
| 40 | * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER |
| 41 | * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, |
| 42 | * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE |
| 43 | * SOFTWARE. |
| 44 | */ |
| 45 | |
| 46 | #include "iblt.hpp" |
| 47 | #include "util.hpp" |
| 48 | |
| 49 | #include <sstream> |
| 50 | |
| 51 | namespace psync { |
| 52 | |
| 53 | const size_t N_HASH(3); |
| 54 | const size_t N_HASHCHECK(11); |
| 55 | |
| 56 | bool |
| 57 | HashTableEntry::isPure() const |
| 58 | { |
| 59 | if (count == 1 || count == -1) { |
| 60 | uint32_t check = murmurHash3(N_HASHCHECK, keySum); |
| 61 | return keyCheck == check; |
| 62 | } |
| 63 | |
| 64 | return false; |
| 65 | } |
| 66 | |
| 67 | bool |
| 68 | HashTableEntry::isEmpty() const |
| 69 | { |
| 70 | return count == 0 && keySum == 0 && keyCheck == 0; |
| 71 | } |
| 72 | |
| 73 | IBLT::IBLT(size_t expectedNumEntries) |
| 74 | { |
| 75 | // 1.5x expectedNumEntries gives very low probability of decoding failure |
| 76 | size_t nEntries = expectedNumEntries + expectedNumEntries / 2; |
| 77 | // make nEntries exactly divisible by N_HASH |
| 78 | size_t remainder = nEntries % N_HASH; |
| 79 | if (remainder != 0) { |
| 80 | nEntries += (N_HASH - remainder); |
| 81 | } |
| 82 | |
| 83 | m_hashTable.resize(nEntries); |
| 84 | } |
| 85 | |
| 86 | void |
| 87 | IBLT::initialize(const ndn::name::Component& ibltName) |
| 88 | { |
| 89 | const auto& values = extractValueFromName(ibltName); |
| 90 | |
| 91 | if (3 * m_hashTable.size() != values.size()) { |
| 92 | BOOST_THROW_EXCEPTION(Error("Received IBF cannot be decoded!")); |
| 93 | } |
| 94 | |
| 95 | for (size_t i = 0; i < m_hashTable.size(); i++) { |
| 96 | HashTableEntry& entry = m_hashTable.at(i); |
| 97 | if (values[i * 3] != 0) { |
| 98 | entry.count = values[i * 3]; |
| 99 | entry.keySum = values[(i * 3) + 1]; |
| 100 | entry.keyCheck = values[(i * 3) + 2]; |
| 101 | } |
| 102 | } |
| 103 | } |
| 104 | |
| 105 | void |
| 106 | IBLT::update(int plusOrMinus, uint32_t key) |
| 107 | { |
| 108 | size_t bucketsPerHash = m_hashTable.size() / N_HASH; |
| 109 | |
| 110 | for (size_t i = 0; i < N_HASH; i++) { |
| 111 | size_t startEntry = i * bucketsPerHash; |
| 112 | uint32_t h = murmurHash3(i, key); |
| 113 | HashTableEntry& entry = m_hashTable.at(startEntry + (h % bucketsPerHash)); |
| 114 | entry.count += plusOrMinus; |
| 115 | entry.keySum ^= key; |
| 116 | entry.keyCheck ^= murmurHash3(N_HASHCHECK, key); |
| 117 | } |
| 118 | } |
| 119 | |
| 120 | void |
| 121 | IBLT::insert(uint32_t key) |
| 122 | { |
| 123 | update(INSERT, key); |
| 124 | } |
| 125 | |
| 126 | void |
| 127 | IBLT::erase(uint32_t key) |
| 128 | { |
| 129 | update(ERASE, key); |
| 130 | } |
| 131 | |
| 132 | bool |
| 133 | IBLT::listEntries(std::set<uint32_t>& positive, std::set<uint32_t>& negative) const |
| 134 | { |
| 135 | IBLT peeled = *this; |
| 136 | |
| 137 | size_t nErased = 0; |
| 138 | do { |
| 139 | nErased = 0; |
| 140 | for (const auto& entry : peeled.m_hashTable) { |
| 141 | if (entry.isPure()) { |
| 142 | if (entry.count == 1) { |
| 143 | positive.insert(entry.keySum); |
| 144 | } |
| 145 | else { |
| 146 | negative.insert(entry.keySum); |
| 147 | } |
| 148 | peeled.update(-entry.count, entry.keySum); |
| 149 | ++nErased; |
| 150 | } |
| 151 | } |
| 152 | } while (nErased > 0); |
| 153 | |
| 154 | // If any buckets for one of the hash functions is not empty, |
| 155 | // then we didn't peel them all: |
| 156 | for (const auto& entry : peeled.m_hashTable) { |
| 157 | if (entry.isEmpty() != true) { |
| 158 | return false; |
| 159 | } |
| 160 | } |
| 161 | |
| 162 | return true; |
| 163 | } |
| 164 | |
| 165 | IBLT |
| 166 | IBLT::operator-(const IBLT& other) const |
| 167 | { |
| 168 | BOOST_ASSERT(m_hashTable.size() == other.m_hashTable.size()); |
| 169 | |
| 170 | IBLT result(*this); |
| 171 | for (size_t i = 0; i < m_hashTable.size(); i++) { |
| 172 | HashTableEntry& e1 = result.m_hashTable.at(i); |
| 173 | const HashTableEntry& e2 = other.m_hashTable.at(i); |
| 174 | e1.count -= e2.count; |
| 175 | e1.keySum ^= e2.keySum; |
| 176 | e1.keyCheck ^= e2.keyCheck; |
| 177 | } |
| 178 | |
| 179 | return result; |
| 180 | } |
| 181 | |
| 182 | bool |
| 183 | operator==(const IBLT& iblt1, const IBLT& iblt2) |
| 184 | { |
| 185 | auto iblt1HashTable = iblt1.getHashTable(); |
| 186 | auto iblt2HashTable = iblt2.getHashTable(); |
| 187 | if (iblt1HashTable.size() != iblt2HashTable.size()) { |
| 188 | return false; |
| 189 | } |
| 190 | |
| 191 | size_t N = iblt1HashTable.size(); |
| 192 | |
| 193 | for (size_t i = 0; i < N; i++) { |
| 194 | if (iblt1HashTable[i].count != iblt2HashTable[i].count || |
| 195 | iblt1HashTable[i].keySum != iblt2HashTable[i].keySum || |
| 196 | iblt1HashTable[i].keyCheck != iblt2HashTable[i].keyCheck) |
| 197 | return false; |
| 198 | } |
| 199 | |
| 200 | return true; |
| 201 | } |
| 202 | |
| 203 | bool |
| 204 | operator!=(const IBLT& iblt1, const IBLT& iblt2) |
| 205 | { |
| 206 | return !(iblt1 == iblt2); |
| 207 | } |
| 208 | |
| 209 | std::ostream& |
| 210 | operator<<(std::ostream& out, const IBLT& iblt) |
| 211 | { |
| 212 | out << "count keySum keyCheckMatch\n"; |
| 213 | for (const auto& entry : iblt.getHashTable()) { |
| 214 | out << entry.count << " " << entry.keySum << " "; |
| 215 | out << ((murmurHash3(N_HASHCHECK, entry.keySum) == entry.keyCheck) || |
| 216 | (entry.isEmpty())? "true" : "false"); |
| 217 | out << "\n"; |
| 218 | } |
| 219 | |
| 220 | return out; |
| 221 | } |
| 222 | |
| 223 | void |
| 224 | IBLT::appendToName(ndn::Name& name) const |
| 225 | { |
| 226 | size_t n = m_hashTable.size(); |
| 227 | size_t unitSize = (32 * 3) / 8; // hard coding |
| 228 | size_t tableSize = unitSize * n; |
| 229 | |
| 230 | std::vector <uint8_t> table(tableSize); |
| 231 | |
| 232 | for (size_t i = 0; i < n; i++) { |
| 233 | // table[i*12], table[i*12+1], table[i*12+2], table[i*12+3] --> hashTable[i].count |
| 234 | |
| 235 | table[(i * unitSize)] = 0xFF & m_hashTable[i].count; |
| 236 | table[(i * unitSize) + 1] = 0xFF & (m_hashTable[i].count >> 8); |
| 237 | table[(i * unitSize) + 2] = 0xFF & (m_hashTable[i].count >> 16); |
| 238 | table[(i * unitSize) + 3] = 0xFF & (m_hashTable[i].count >> 24); |
| 239 | |
| 240 | // table[i*12+4], table[i*12+5], table[i*12+6], table[i*12+7] --> hashTable[i].keySum |
| 241 | |
| 242 | table[(i * unitSize) + 4] = 0xFF & m_hashTable[i].keySum; |
| 243 | table[(i * unitSize) + 5] = 0xFF & (m_hashTable[i].keySum >> 8); |
| 244 | table[(i * unitSize) + 6] = 0xFF & (m_hashTable[i].keySum >> 16); |
| 245 | table[(i * unitSize) + 7] = 0xFF & (m_hashTable[i].keySum >> 24); |
| 246 | |
| 247 | // table[i*12+8], table[i*12+9], table[i*12+10], table[i*12+11] --> hashTable[i].keyCheck |
| 248 | |
| 249 | table[(i * unitSize) + 8] = 0xFF & m_hashTable[i].keyCheck; |
| 250 | table[(i * unitSize) + 9] = 0xFF & (m_hashTable[i].keyCheck >> 8); |
| 251 | table[(i * unitSize) + 10] = 0xFF & (m_hashTable[i].keyCheck >> 16); |
| 252 | table[(i * unitSize) + 11] = 0xFF & (m_hashTable[i].keyCheck >> 24); |
| 253 | } |
| 254 | |
| 255 | name.append(table.begin(), table.end()); |
| 256 | } |
| 257 | |
| 258 | std::vector<uint32_t> |
| 259 | IBLT::extractValueFromName(const ndn::name::Component& ibltName) const |
| 260 | { |
| 261 | std::vector<uint8_t> ibltValues(ibltName.value_begin(), ibltName.value_end()); |
| 262 | size_t n = ibltValues.size() / 4; |
| 263 | |
| 264 | std::vector<uint32_t> values(n, 0); |
| 265 | |
| 266 | for (size_t i = 0; i < 4 * n; i += 4) { |
| 267 | uint32_t t = (ibltValues[i + 3] << 24) + |
| 268 | (ibltValues[i + 2] << 16) + |
| 269 | (ibltValues[i + 1] << 8) + |
| 270 | ibltValues[i]; |
| 271 | values[i / 4] = t; |
| 272 | } |
| 273 | |
| 274 | return values; |
| 275 | } |
| 276 | |
| 277 | } // namespace psync |