includes/libs/HashRing.php

   1 <?php
   2 /**
   3  * Convenience class for weighted consistent hash rings.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 /**
  24  * Convenience class for weighted consistent hash rings
  25  *
  26  * This deterministically maps "keys" to a set of "locations" while avoiding clumping
  27  *
  28  * Each location is represented by a number of nodes on a ring proportionate to the ratio
  29  * of its weight compared to the total location weight. Note positions are deterministically
  30  * derived from the hash of the location name. Nodes are responsible for the portion of the
  31  * ring, counter-clockwise, up until the next node. Locations are responsible for all portions
  32  * of the ring that the location's nodes are responsible for.
  33  *
  34  * A location that is temporarily "ejected" is said to be absent from the "live" ring.
  35  * If no location ejections are active, then the base ring and live ring are identical.
  36  *
  37  * @since 1.22
  38  */
  39 class HashRing implements Serializable {
  40         /** @var string Hashing algorithm for hash() */
  41         protected $algo;
  42         /** @var int[] Non-empty (location => integer weight) */
  43         protected $weightByLocation;
  44         /** @var int[] Map of (location => UNIX timestamp) */
  45         protected $ejectExpiryByLocation;
  46
  47         /** @var array[] Non-empty list of (float, node name, location name) */
  48         protected $baseRing;
  49         /** @var array[] Non-empty list of (float, node name, location name) */
  50         protected $liveRing;
  51
  52         /** @var float Number of positions on the ring */
  53         const RING_SIZE = 4294967296.0; // 2^32
  54         /** @var integer Overall number of node groups per server */
  55         const HASHES_PER_LOCATION = 40;
  56         /** @var integer Number of nodes in a node group */
  57         const SECTORS_PER_HASH = 4;
  58
  59         const KEY_POS = 0;
  60         const KEY_LOCATION = 1;
  61
  62         /** @var int Consider all locations */
  63         const RING_ALL = 0;
  64         /** @var int Only consider "live" locations */
  65         const RING_LIVE = 1;
  66
  67         /**
  68          * Make a consistent hash ring given a set of locations and their weight values
  69          *
  70          * @param int[] $map Map of (location => weight)
  71          * @param string $algo Hashing algorithm listed in hash_algos() [optional]
  72          * @param int[] $ejections Map of (location => UNIX timestamp) for ejection expiries
  73          * @since 1.31
  74          */
  75         public function __construct( array $map, $algo = 'sha1', array $ejections = [] ) {
  76                 $this->init( $map, $algo, $ejections );
  77         }
  78
  79         /**
  80          * @param int[] $map Map of (location => integer)
  81          * @param string $algo Hashing algorithm
  82          * @param int[] $ejections Map of (location => UNIX timestamp) for ejection expires
  83          */
  84         protected function init( array $map, $algo, array $ejections ) {
  85                 if ( !in_array( $algo, hash_algos(), true ) ) {
  86                         throw new RuntimeException( __METHOD__ . ": unsupported '$algo' hash algorithm." );
  87                 }
  88
  89                 $weightByLocation = array_filter( $map );
  90                 if ( $weightByLocation === [] ) {
  91                         throw new UnexpectedValueException( "No locations with non-zero weight." );
  92                 } elseif ( min( $map ) < 0 ) {
  93                         throw new InvalidArgumentException( "Location weight cannot be negative." );
  94                 }
  95
  96                 $this->algo = $algo;
  97                 $this->weightByLocation = $weightByLocation;
  98                 $this->ejectExpiryByLocation = $ejections;
  99                 $this->baseRing = $this->buildLocationRing( $this->weightByLocation, $this->algo );
 100         }
 101
 102         /**
 103          * Get the location of an item on the ring
 104          *
 105          * @param string $item
 106          * @return string Location
 107          * @throws UnexpectedValueException
 108          */
 109         final public function getLocation( $item ) {
 110                 return $this->getLocations( $item, 1 )[0];
 111         }
 112
 113         /**
 114          * Get the location of an item on the ring, as well as the next locations
 115          *
 116          * @param string $item
 117          * @param int $limit Maximum number of locations to return
 118          * @param int $from One of the RING_* class constants
 119          * @return string[] List of locations
 120          * @throws UnexpectedValueException
 121          */
 122         public function getLocations( $item, $limit, $from = self::RING_ALL ) {
 123                 if ( $from === self::RING_ALL ) {
 124                         $ring = $this->baseRing;
 125                 } elseif ( $from === self::RING_LIVE ) {
 126                         $ring = $this->getLiveRing();
 127                 } else {
 128                         throw new InvalidArgumentException( "Invalid ring source specified." );
 129                 }
 130
 131                 // Locate this item's position on the hash ring
 132                 $position = $this->getItemPosition( $item );
 133                 $itemNodeIndex = $this->findNodeIndexForPosition( $position, $ring );
 134
 135                 $locations = [];
 136                 $currentIndex = $itemNodeIndex;
 137                 while ( count( $locations ) < $limit ) {
 138                         $nodeLocation = $ring[$currentIndex][self::KEY_LOCATION];
 139                         if ( !in_array( $nodeLocation, $locations, true ) ) {
 140                                 // Ignore other nodes for the same locations already added
 141                                 $locations[] = $nodeLocation;
 142                         }
 143                         $currentIndex = $this->getNextClockwiseNodeIndex( $currentIndex, $ring );
 144                         if ( $currentIndex === $itemNodeIndex ) {
 145                                 break; // all nodes visited
 146                         }
 147                 }
 148
 149                 return $locations;
 150         }
 151
 152         /**
 153          * @param float $position
 154          * @param array[] $ring Either the base or live ring
 155          * @return int|null
 156          */
 157         private function findNodeIndexForPosition( $position, $ring ) {
 158                 $count = count( $ring );
 159                 if ( $count === 0 ) {
 160                         return null;
 161                 }
 162                 $lowPos = 0;
 163                 $highPos = $count;
 164                 while ( true ) {
 165                         $midPos = intval( ( $lowPos + $highPos ) / 2 );
 166                         if ( $midPos === $count ) {
 167                                 return 0;
 168                         }
 169                         $midVal = $ring[$midPos][self::KEY_POS];
 170                         $midMinusOneVal = $midPos === 0 ? 0 : $ring[$midPos - 1][self::KEY_POS];
 171
 172                         if ( $position <= $midVal && $position > $midMinusOneVal ) {
 173                                 return $midPos;
 174                         }
 175
 176                         if ( $midVal < $position ) {
 177                                 $lowPos = $midPos + 1;
 178                         } else {
 179                                 $highPos = $midPos - 1;
 180                         }
 181
 182                         if ( $lowPos > $highPos ) {
 183                                 return 0;
 184                         }
 185                 }
 186         }
 187
 188         /**
 189          * Get the map of locations to weight (does not include zero weight items)
 190          *
 191          * @return int[]
 192          */
 193         public function getLocationWeights() {
 194                 return $this->weightByLocation;
 195         }
 196
 197         /**
 198          * Remove a location from the "live" hash ring
 199          *
 200          * @param string $location
 201          * @param int $ttl Seconds
 202          * @return bool Whether some non-ejected locations are left
 203          * @throws UnexpectedValueException
 204          */
 205         public function ejectFromLiveRing( $location, $ttl ) {
 206                 if ( !isset( $this->weightByLocation[$location] ) ) {
 207                         throw new UnexpectedValueException( "No location '$location' in the ring." );
 208                 }
 209
 210                 $expiry = $this->getCurrentTime() + $ttl;
 211                 $this->ejectExpiryByLocation[$location] = $expiry;
 212
 213                 $this->liveRing = null; // invalidate ring cache
 214
 215                 return ( count( $this->ejectExpiryByLocation ) < count( $this->weightByLocation ) );
 216         }
 217
 218         /**
 219          * Get the location of an item on the "live" ring
 220          *
 221          * @param string $item
 222          * @return string Location
 223          * @throws UnexpectedValueException
 224          */
 225         final public function getLiveLocation( $item ) {
 226                 return $this->getLocations( $item, 1, self::RING_LIVE )[0];
 227         }
 228
 229         /**
 230          * Get the location of an item on the "live" ring, as well as the next locations
 231          *
 232          * @param string $item
 233          * @param int $limit Maximum number of locations to return
 234          * @return string[] List of locations
 235          * @throws UnexpectedValueException
 236          */
 237         final public function getLiveLocations( $item, $limit ) {
 238                 return $this->getLocations( $item, $limit, self::RING_LIVE );
 239         }
 240
 241         /**
 242          * Get the map of "live" locations to weight (does not include zero weight items)
 243          *
 244          * @return int[]
 245          * @throws UnexpectedValueException
 246          */
 247         public function getLiveLocationWeights() {
 248                 $now = $this->getCurrentTime();
 249
 250                 return array_diff_key(
 251                         $this->weightByLocation,
 252                         array_filter(
 253                                 $this->ejectExpiryByLocation,
 254                                 function ( $expiry ) use ( $now ) {
 255                                         return ( $expiry > $now );
 256                                 }
 257                         )
 258                 );
 259         }
 260
 261         /**
 262          * @param int[] $weightByLocation
 263          * @param string $algo Hashing algorithm
 264          * @return array[]
 265          */
 266         private function buildLocationRing( array $weightByLocation, $algo ) {
 267                 $locationCount = count( $weightByLocation );
 268                 $totalWeight = array_sum( $weightByLocation );
 269
 270                 $ring = [];
 271                 // Assign nodes to all locations based on location weight
 272                 $claimed = []; // (position as string => (node, index))
 273                 foreach ( $weightByLocation as $location => $weight ) {
 274                         $ratio = $weight / $totalWeight;
 275                         // There $locationCount * (HASHES_PER_LOCATION * 4) nodes available;
 276                         // assign a few groups of nodes to this location based on its weight.
 277                         $nodesQuartets = intval( $ratio * self::HASHES_PER_LOCATION * $locationCount );
 278                         for ( $qi = 0; $qi < $nodesQuartets; ++$qi ) {
 279                                 // For efficiency, get 4 points per hash call and 4X node count.
 280                                 // If $algo is MD5, then this matches that of with libketama.
 281                                 // See https://github.com/RJ/ketama/blob/master/libketama/ketama.c
 282                                 $positions = $this->getNodePositionQuartet( "{$location}-{$qi}" );
 283                                 foreach ( $positions as $gi => $position ) {
 284                                         $node = ( $qi * self::SECTORS_PER_HASH + $gi ) . "@$location";
 285                                         $posKey = (string)$position; // large integer
 286                                         if ( isset( $claimed[$posKey] ) ) {
 287                                                 // Disallow duplicates for sanity (name decides precedence)
 288                                                 if ( $claimed[$posKey]['node'] > $node ) {
 289                                                         continue;
 290                                                 } else {
 291                                                         unset( $ring[$claimed[$posKey]['index']] );
 292                                                 }
 293                                         }
 294                                         $ring[] = [
 295                                                 self::KEY_POS => $position,
 296                                                 self::KEY_LOCATION => $location
 297                                         ];
 298                                         $claimed[$posKey] = [ 'node' => $node, 'index' => count( $ring ) - 1 ];
 299                                 }
 300                         }
 301                 }
 302                 // Sort the locations into clockwise order based on the hash ring position
 303                 usort( $ring, function ( $a, $b ) {
 304                         if ( $a[self::KEY_POS] === $b[self::KEY_POS] ) {
 305                                 throw new UnexpectedValueException( 'Duplicate node positions.' );
 306                         }
 307
 308                         return ( $a[self::KEY_POS] < $b[self::KEY_POS] ? -1 : 1 );
 309                 } );
 310
 311                 return $ring;
 312         }
 313
 314         /**
 315          * @param string $item Key
 316          * @return float Ring position; integral number in [0, self::RING_SIZE - 1]
 317          */
 318         private function getItemPosition( $item ) {
 319                 // If $algo is MD5, then this matches that of with libketama.
 320                 // See https://github.com/RJ/ketama/blob/master/libketama/ketama.c
 321                 $octets = substr( hash( $this->algo, (string)$item, true ), 0, 4 );
 322                 if ( strlen( $octets ) != 4 ) {
 323                         throw new UnexpectedValueException( __METHOD__ . ": {$this->algo} is < 32 bits." );
 324                 }
 325
 326                 return (float)sprintf( '%u', unpack( 'V', $octets )[1] );
 327         }
 328
 329         /**
 330          * @param string $nodeGroupName
 331          * @return float[] Four ring positions on [0, self::RING_SIZE - 1]
 332          */
 333         private function getNodePositionQuartet( $nodeGroupName ) {
 334                 $octets = substr( hash( $this->algo, (string)$nodeGroupName, true ), 0, 16 );
 335                 if ( strlen( $octets ) != 16 ) {
 336                         throw new UnexpectedValueException( __METHOD__ . ": {$this->algo} is < 128 bits." );
 337                 }
 338
 339                 $positions = [];
 340                 foreach ( unpack( 'V4', $octets ) as $signed ) {
 341                         $positions[] = (float)sprintf( '%u', $signed );
 342                 }
 343
 344                 return $positions;
 345         }
 346
 347         /**
 348          * @param int $i Valid index for a node in the ring
 349          * @param array[] $ring Either the base or live ring
 350          * @return int Valid index for a node in the ring
 351          */
 352         private function getNextClockwiseNodeIndex( $i, $ring ) {
 353                 if ( !isset( $ring[$i] ) ) {
 354                         throw new UnexpectedValueException( __METHOD__ . ": reference index is invalid." );
 355                 }
 356
 357                 $next = $i + 1;
 358
 359                 return ( $next < count( $ring ) ) ? $next : 0;
 360         }
 361
 362         /**
 363          * Get the "live" hash ring (which does not include ejected locations)
 364          *
 365          * @return array[]
 366          * @throws UnexpectedValueException
 367          */
 368         protected function getLiveRing() {
 369                 if ( !$this->ejectExpiryByLocation ) {
 370                         return $this->baseRing; // nothing ejected
 371                 }
 372
 373                 $now = $this->getCurrentTime();
 374
 375                 if ( $this->liveRing === null || min( $this->ejectExpiryByLocation ) <= $now ) {
 376                         // Live ring needs to be regerenated...
 377                         $this->ejectExpiryByLocation = array_filter(
 378                                 $this->ejectExpiryByLocation,
 379                                 function ( $expiry ) use ( $now ) {
 380                                         return ( $expiry > $now );
 381                                 }
 382                         );
 383
 384                         if ( count( $this->ejectExpiryByLocation ) ) {
 385                                 // Some locations are still ejected from the ring
 386                                 $liveRing = [];
 387                                 foreach ( $this->baseRing as $i => $nodeInfo ) {
 388                                         $location = $nodeInfo[self::KEY_LOCATION];
 389                                         if ( !isset( $this->ejectExpiryByLocation[$location] ) ) {
 390                                                 $liveRing[] = $nodeInfo;
 391                                         }
 392                                 }
 393                         } else {
 394                                 $liveRing = $this->baseRing;
 395                         }
 396
 397                         $this->liveRing = $liveRing;
 398                 }
 399
 400                 if ( !$this->liveRing ) {
 401                         throw new UnexpectedValueException( "The live ring is currently empty." );
 402                 }
 403
 404                 return $this->liveRing;
 405         }
 406
 407         /**
 408          * @return int UNIX timestamp
 409          */
 410         protected function getCurrentTime() {
 411                 return time();
 412         }
 413
 414         public function serialize() {
 415                 return serialize( [
 416                         'algorithm' => $this->algo,
 417                         'locations' => $this->weightByLocation,
 418                         'ejections' => $this->ejectExpiryByLocation
 419                 ] );
 420         }
 421
 422         public function unserialize( $serialized ) {
 423                 $data = unserialize( $serialized );
 424                 if ( is_array( $data ) ) {
 425                         $this->init( $data['locations'], $data['algorithm'], $data['ejections'] );
 426                 } else {
 427                         throw new UnexpectedValueException( __METHOD__ . ": unable to decode JSON." );
 428                 }
 429         }
 430 }