1. ã½ã¼ã·ã£ã«ã²ã¼ã ã¹ã±ã¼ã«ã¢ã¦ãã®æ´å² gussan@Drecom Co., Ltd. Copyright © Drecom Co., Ltd. 12å¹´2æ20æ¥æææ¥
Facebookã®æ°ãããªã¢ã«ã¿ã¤ã 解æã®ã·ã¹ãã ã§ã¯ãHBaseã§1æ¥200å件ã®ã¤ãã³ããå¦çãã¦ããããã§ãã以ä¸ã®è¨äºã®ç¿»è¨³ã§ããHigh Scalability - High Scalability - Facebookâs New Realtime Analytics System: HBase to Process 20 Billion Events Per DayFacebookãã¾ããã£ã¦ããããå½¼ãã¯å·¨å¤§ãªãªã¢ã«ã¿ã¤ã ãã¼ã¿ã®ã¹ããªã¼ã ãå¦çãããã1ã¤ã®ã·ã¹ãã ãæ§ç¯ããã®ã ã以åã«ãFacebookã¯ãªã¢ã«ã¿ã¤ã ãªã¡ãã»ã¼ã¸ã·ã¹ãã ãHBaseã§æ§ç¯ãã¦ãã(http://highscalability.com/blog/2010/11/16/facebooks-new-real-time-messaging-system-hbase-to-store-135.ht
Twitterã®å¤§è¦æ¨¡ã·ã¹ãã éç¨æè¡ããããã¯ã¯ã¸ã©ã®è ¹ã®ä¸ï¼åç·¨ï¼ï½ãã°ã®ç§å¦çãªåæã¨ãTwitterã®ããã¼ã¯ã¢ã¼ãã å é±ã®6æ22æ¥ãããç±³ãµã³ã¿ã¯ã©ã©ã§è¡ããã¦ããWebãµã¤ãã®ããã©ã¼ãã³ã¹ã¨éç¨ã«é¢ãããªã©ã¤ãªã¼ã®ã¤ãã³ããVelocity 2010ããéå¬ããã¦ãã¾ããã ãã®ä¸ã§ãTwitterã®John Adamsæ°ãTwitterã®ã·ã¹ãã éç¨ã«ã¤ãã¦èª¬æããã»ãã·ã§ã³ãIn the Belly of the Whale: Operations at Twitterãï¼ã¯ã¸ã©ã®è ¹ã®ä¸ï¼Twitterã§ã®éç¨ï¼ãè¡ããã¦ãã¾ããTwitterã®ãããªå¤§è¦æ¨¡ãã¤ãªã¢ã«ã¿ã¤ã ãªWebãµã¤ãã®éç¨ã¨ã¯ã©ããããã®ãªã®ã§ããããï¼ å ¬éããã¦ããã»ãã·ã§ã³ã®å 容ãåºã«æ¦è¦ãè¨äºã§ç´¹ä»ãã¾ããããã·ã¹ãã 管çè ã®æ°ããªå½¹å²ãRailsã®æ§è½ã®è©ä¾¡ãBittorrentã使ã£ã
å æãNot Only NoSQL!! é©ç°çãªã¾ã§ã«WRITEæ§è½ãã¹ã±ã¼ã«ãããSPIDERã¹ãã¬ã¼ã¸ã¨ã³ã¸ã³ã¨ããã¨ã³ããªã§SPIDERã¹ãã¬ã¼ã¸ã¨ã³ã¸ã³ã«ããã¹ã±ã¼ã«ã¢ã¦ããåãï¼ã¨ãã話ãæ¸ãããSPIDERã¹ãã¬ã¼ã¸ã¨ã³ã¸ã³ã¯åãã¤ãã ãããã¦ãã¦ããã¾ãã¦ã§ãä¸ã§è¦ã¤ãããªããå¯ä¸è¦ã¤ããæ¥æ¬èªã®è¨äºã¯ãã¦ãã¦ã©ãã«ãããå½ç£MySQLã¹ãã¬ã¼ã¸ã¨ã³ã¸ã³ãSpiderãã®ä½è ãæ¯æ³¢å¥å¾³æ°ã«èã ãã ãã§ãããSPIDERã¹ãã¬ã¼ã¸ã¨ã³ã¸ã³ã¯æ¯æ³¢æ°ã«ããåç¬ã®ä½åã§ãããããæ¯æ³¢æ°ã¯éçºã ãã§æãã£ã±ãã§ããã使ãæ¹ã®ç´¹ä»è¨äºãæ¸ããã¨ã¾ã§ã¯æãåããªãã®ã§ãããããããªåãã¹ãã¬ã¼ã¸ã¨ã³ã¸ã³ãããã¥ã¡ã³ãã足ããªãããã§ä½¿ã£ã¦è²°ããªããªãã¦å¿ä½ãªãï¼ï¼ ã¨ããããã§ãä»æ¥ã¯SPIDERã¹ãã¬ã¼ã¸ã¨ã³ã¸ã³ã®åºæ¬çãªä½¿ãæ¹ã«ã¤ãã¦ç´¹ä»ãããå°ãé·ãã¨ã³ããªã§ããããæå¾ã¾ã§ãä»ã
Twitterãåæ£ãã¬ã¼ã ã¯ã¼ã¯ãGizzardãå ¬éï¼ Scalaã§æ¸ãããShardingãå®ç¾ããããã«ã¦ã§ã¢ Twitterã¯ç¬èªã«éçºããåæ£ãã¬ã¼ã ã¯ã¼ã¯ã®ãGizzardãããªã¼ãã³ã½ã¼ã¹ã¨ãã¦å ¬éãã¾ãããGizzardã¯Scalaã§æ¸ãããJavaVMä¸ã§åä½ããããã«ã¦ã§ã¢ã§ãPHPãRubyã¨ãã£ãWebã¢ããªã±ã¼ã·ã§ã³ããã®è¦æ±ãèªåçã«ãã¼ã¿ãã¼ã¹ã«åæ£ãããã¨ã§ã大è¦æ¨¡ã§å¯ç¨æ§ã®é«ãåæ£ãã¼ã¿ãã¼ã¹ã容æã«å®ç¾ããããã®ãã®ã§ãã Gizzardï¼ãã©ã«ããã¬ã©ã³ããªåæ£ãã¼ã¿ãã¼ã¹ãå®ç¾ The Twitter Engineering Blog: Introducing Gizzard, a framework for creating distributed datastores Twitterã®ããã°ã«ãã¹ãããããIntroducing Gizzard
scale outã®æè¡ é¦è¤ ä¸å¹¸ Last-updated: January 5, 2010 注: ãã®ãã¼ã¸ã®æç« ã¯ä»¥ä¸ã®è¨äºã®å å稿ã§ãã é¦è¤ä¸å¹¸, "ã¹ã±ã¼ã«ã¢ã¦ãã®æè¡", ã¯ã©ã¦ãã®æè¡, pp.88-101, (æ ª)ã¢ã¹ãã¼ã»ã¡ãã£ã¢ã¯ã¼ã¯ã¹, ISBN978-4-04-868064-6, 2009å¹´ 11æ 6æ¥ ã¢ã¹ãã¼ã»ã¡ãã£ã¢ã¯ã¼ã¯ã¹ç¤¾ã® æ¸ç±ç´¹ä»ãã¼ã¸ Amazon.co.jp ã® ãã¼ã¸ é¦è¤ä¸å¹¸, "ã¹ã±ã¼ã«ã¢ã¦ãã®æè¡", UNIX magazine 2009å¹´ 4æå·, pp.78-91, (æ ª)ã¢ã¹ãã¼ã»ã¡ãã£ã¢ã¯ã¼ã¯ã¹, 2009å¹´ 3æ 18æ¥ ãã¼ã¿ãã¼ã¹ã«æ±ããããæ§è½ã試ç®ããã¨ããã åå°ãç¾å°â¦æ°ä¸å°ã®ãµã¼ããå¿ è¦ã«ãªã£ãã ã¯ã©ã¦ããæ§ç¯ããå´ã¯ããããåé¡ã«ç´é¢ãã解決ãããã¨ãã¦ããã å°æ°ã«æ¯ä¾ããæ§è½ãå¼ãåºããã¨ãã¤ã¾ãsca
KOF2009ã«ã¦ããã¦ã§ããµã¼ãã¹ã®ããã©ã¼ãã³ã¹ã¨ã¹ã±ã¼ã©ããªãã£ãã¨é¡ãã¦çºè¡¨ãã¦ãã¾ãããçºè¡¨è³æã以ä¸ã«ç½®ãã¦ããã¾ãã Performance and Scalability of Web ServiceView more presentations from Shinji Tanaka. æ¦è¦ã¯ããã¦ã§ããµã¼ãã¹ã®ããã©ã¼ãã³ã¹ãåä¸ããã¹ã±ã¼ã©ããªãã£ãé«ããããã«ãã¯ã¦ãªã§ã¯æ§ã ãªåçµã¿ãè¡ã£ã¦ãã¾ããæ¬ã»ãã·ã§ã³ã§ã¯ãã¯ã¦ãªã§æ¡ç¨ãã¦ããå ·ä½çãªæè¡ããã¦ãã¦ãå¯è¦åææ³ã¨ããããã®å¹æã«ã¤ãã¦ç´¹ä»ãã¾ãããã¨ãããã®ã§ãã æè¿ã®ãInteropãã«ã¼ãã«èªæ¸ä¼ãããã§è©±ããå 容ãã¾ã¨ãã¤ã¤ãã¬ã¹ãã³ã¹ã¿ã¤ã ã®å¯è¦åã¨ããæè¿ã®åãçµã¿ã«ã¤ãã¦è©±ãã¾ããã æè¿ãã¬ã¹ãã³ã¹ã¿ã¤ã ã«ã¤ãã¦ã¯ã以ä¸ã®ãããªã°ã©ãã使ã£ã¦ãã¾ãã x軸ãã¬ã¹ãã³ã¹æéãy軸ããã®æéå ã«å
Facebookã大è¦æ¨¡ã¹ã±ã¼ã©ããªãã£ã¸ã®ææ¦ã§å¦ãã ãã¨ï¼å¾ç·¨ï¼ï½ãã£ãã·ã¥ãæ±ããã¹ã±ã¼ã©ããªãã£ã®åé¡ã¨ãã¼ã¿ã»ã³ã¿ã¼ã«ã¾ãããä¸è²«æ§ å ¨ä¸çã§3å人ãè¶ ããä¼å¡ãæ±ããä¸çæ大ã®SNSã¨ãªã£ãFacebookãå社ã®æè¡æ å½ãã¤ã¹ãã¬ã¸ãã³ã Jeff Rothschildæ°ãã10æ8æ¥ã«ç±³ã«ãªãã©ã«ãã¢å¤§å¦ãµã³ãã£ã¨ã´æ ¡ã§è¡ã£ãã»ããã¼ãHigh Performance at Massive Scale-Lessons learned at Facebookãã®å 容ãåæ§æãã¦ç´¹ä»ãã¾ãã ï¼ãã®è¨äºã¯ãFacebookã大è¦æ¨¡ãªã¹ã±ã¼ã©ããªãã£ã¸ã®ææ¦ã§å¦ãã ãã¨ï¼åç·¨ï¼ï½800åæã®åçãã¼ã¿ã¨PHPã®ã¹ã±ã¼ã©ããªãã£åé¡ãã®ç¶ãã§ãï¼ ãã£ãã·ã¥ãã¹ã±ã¼ã©ããªãã£ã«å¤§ããªå½¹å²ãæããã¦ãã Facebookã®ä¸»ãªå½¹å²ã¯ãã¦ã¼ã¶ã¼ãç°¡åã«ï¼å人ãã¡ã®ï¼æ å ±ãéãããã¨ãã§
Facebookã大è¦æ¨¡ã¹ã±ã¼ã©ããªãã£ã¸ã®ææ¦ã§å¦ãã ãã¨ï¼åç·¨ï¼ï½800åæã®åçãã¼ã¿ã¨PHPã®ã¹ã±ã¼ã©ããªãã£åé¡ å ¨ä¸çã§3å人ãè¶ ããä¼å¡ãæ±ããä¸çæ大ã®SNSã¨ãªã£ãFacebookãå社ã®å·¨å¤§ãªã·ã¹ãã ã¯ã3ã¤ã®ãã¼ã¿ã»ã³ã¿ã¼ã«ããç´3ä¸å°ã®ãµã¼ãã¨ãPHPãC++ãMemcacheãMySQLãªã©ã®ã½ããã¦ã§ã¢ç¾¤ã«ãã£ã¦æ¯ãããã¦ãã¾ãï¼å社ã®ãã¼ã¿ã»ã³ã¿ã¼ã®å·¨å¤§ãã¯ãè¨äºã3åã®ã¦ã¼ã¶ã¼ãæ±ããFacebookã®ãã¼ã¿ã»ã³ã¿ã¼ã移åã¯èªè»¢è»ãå¸æã¯100Gbã¤ã¼ãµããã ããåç §ï¼ã å社ã®æè¡æ å½ãã¤ã¹ãã¬ã¸ãã³ã Jeff Rothschildæ°ã¯ãFacebookãå®ç¾ãã¦ãã大è¦æ¨¡ãªã¹ã±ã¼ã©ããªãã£ããããã«ãã¦ãããã®ã½ããã¦ã§ã¢ã§å®ç¾ãã¦ããã®ãã10æ8æ¥ã«ç±³ã«ãªãã©ã«ãã¢å¤§å¦ãµã³ãã£ã¨ã´æ ¡ã§è¡ã£ãã»ããã¼ãHigh Performance at Mas
å é±éææ¥ãBPStudy#25ã§ããããã©ã¼ãã³ã¹ã¨ã¹ã±ã¼ã©ããªãã£ã®ããã®ãã¼ã¿ãã¼ã¹ã¢ã¼ããã¯ãã£ãã¨ããé¡ç®ã§è©±ãããã¦ããã ãã¾ããããã®éã«ä½¿ç¨ããçºè¡¨è³æã¯ä»¥ä¸ã®ã¨ããã§ãã 1. Happy Optimization æåã«ãæé©åã®èãæ¹ã¨ãã¦ãä¸éå¤ãäºæ¸¬ãããããå ã«ãªã½ã¼ã¹é åãèãããã¨ããææ³ã説æãã¾ããã
å°ç¨ãµã¼ãã®ã¯ããªã®ã«ãåå¥ã¨ã³ããªã¼ã®åæ§ç¯ã ãã§7æé以ä¸ãããã£ã¦ã©ããªãã ããâ¥â¥ã¨æã£ã¦ããã¨ãããéæ³ä½¿ããç¾ãã¦ãã¥ã¼ãã³ã°ãã¦ããã¾ããï¼ ãã¤ãã¿ã¼ã§æ²ããã«ã¤ã¶ããã¦ããã¨ãããéæ³ä½¿ããç¾ãã¾ããããã¥ã¼ãã³ã°ãã¦ãããã®ã¯ã@yagishitaï¼Six Apartï¼ã§ãã ãã¿ãã«ããã¥ã¼ãã³ã°ããçµæã¯ããã°ã§å ¬éãã¦ããã²ä»ã®çããã®å½¹ã«ç«ã£ãããããªã¨ãããã¨ã§ããæä¼ããã¦é ãã次第ã§ãã ã¨ãããã¨ã§ä»åãå¥å¹ããã¬ã·ãã¯ãããªæãã§ãã ã»ã¹ãã¬ã¼ã¸ã¨ã³ã¸ã³ãMyISAMããInnoDBã«å¤æ´ãã ã»InnoDBã®ãããã¡ãã¼ã«ã®ãã£ãã·ã¥çãé«ããããã«my.cnfã®è¨å® (innodb_buffer_pool_size) ãå¤æ´ãã ãã¹ãã¬ã¼ã¸ã¨ã³ã¸ã³ãMyISAMããInnoDBã«å¤æ´ãããã®ã大ããã¿ããã§ããã ã¹ãã¬ã¼ã¸ã¨ã³ã¸ã³ã¨ããã®ã¯ã
ã¯ã¦ãªã°ã«ã¼ãã®çµäºæ¥ã2020å¹´1æ31æ¥(é)ã«æ±ºå®ãã¾ãã 以ä¸ã®ã¨ã³ããªã®éããä»å¹´æ«ãç®å¦ã«ã¯ã¦ãªã°ã«ã¼ããçµäºäºå®ã§ããæ¨ããç¥ãããã¦ããã¾ããã 2019å¹´æ«ãç®å¦ã«ãã¯ã¦ãªã°ã«ã¼ãã®æä¾ãçµäºããäºå®ã§ã - ã¯ã¦ãªã°ã«ã¼ãæ¥è¨ ãã®ãã³ãæ£å¼ã«çµäºæ¥ã決å®ãããã¾ããã®ã§ã以ä¸ã®éãã確èªãã ããã çµäºæ¥: 2020å¹´1æ31æ¥(é) ã¨ã¯ã¹ãã¼ãå¸æç³è«æé:2020å¹´1æ31æ¥(é) çµäºæ¥ä»¥éã¯ãã¯ã¦ãªã°ã«ã¼ãã®é²è¦§ããã³æ稿ã¯è¡ãã¾ãããæ¥è¨ã®ã¨ã¯ã¹ãã¼ããå¿ è¦ãªæ¹ã¯ä»¥ä¸ã®è¨äºã«ãããã£ã¦æç¶ãããã¦ãã ããã ã¯ã¦ãªã°ã«ã¼ãã«æ稿ãããæ¥è¨ãã¼ã¿ã®ã¨ã¯ã¹ãã¼ãã«ã¤ã㦠- ã¯ã¦ãªã°ã«ã¼ãæ¥è¨ ãå©ç¨ã®ã¿ãªãã¾ã«ã¯ãè¿·æãããããããã¾ãããã©ãããããããé¡ããããã¾ãã 2020-06-25 è¿½è¨ ã¯ã¦ãªã°ã«ã¼ãæ¥è¨ã®ã¨ã¯ã¹ãã¼ããã¼ã¿ã¯2020å¹´2æ28
ãµã¼ããå®å ¨ã«éç¨ããæ½è¨ã¨ãã¦æ§ç¯ããããã¼ã¿ã»ã³ã¿ã¼ã§ãããã°ã¼ã°ã«ã§ã¯ãã®ãã¼ã¿ã»ã³ã¿ã¼ã§ãã"è½ã¡ã"ãã¨ãããã¨æ³å®ãã¦ã¢ã¼ããã¯ãã£ãæ§ç¯ãã¦ãã¾ãã ç±³ã°ã¼ã°ã«ãä»å¹´ã®5æã«è¡ã£ãã¤ãã³ããGoogle I/Oãã§ãå社ã®Google App Engine datastore leadã§ããRyan Barettæ°ãè¡ã£ãè¬æ¼ãTransactions Across Datacenters (and Other Weekend Projects)ãã®ãããªãYouTubeã§å ¬éããã¾ããã Barettæ°ã¯ãæ å½ãã¦ããGoogle App Engineã®ãã¼ã¿ãã¼ã¹ã«é¢ãã¦ã°ã¼ã°ã«ããmultihomingãï¼ãã«ããã¼ãã³ã°ï¼ã¨å¼ã¶è¤æ°ã®ãã¼ã¿ã»ã³ã¿ã¼ãç¨ããå¦çãå®ç¾ãã¦ããçç±ã¨ãã¦ããã¼ã¿ã»ã³ã¿ã¼ãèªç¶ç½å®³ãåé»ã«è¦èãããããã¡ã³ããã³ã¹ãªã©ã«ãããã¼ã¿ã»ã³ã¿ã¼ã®
å³è§£æ±ãã 以ä¸ããããã³ã«å¦çãã¨ãã¡ãã»ã¼ã¸å¦çããåãã¦æ±ã£ã¦ãããããã®å·®ãé¡èã«åºãã®ã¯å ¨ææ¤ç´¢ã¨ã³ã¸ã³ãéåæã¸ã§ããµã¼ãã¼ãªã©ãå°ããªã¡ãã»ã¼ã¸ã§éãå¦çãããã¿ã¤ããã¹ããªã¼ã æåã®ãããã³ã«ã®å ´åã¯ããããã³ã«å¦çãããã¹ããªã¼ã å¦çãã«ç½®ãæããã¨ãããããããªãã ã·ã³ã°ã«ã¹ã¬ããï½¥ã¤ãã³ãé§å ã³ãã¯ã·ã§ã³Nï¼ã¹ã¬ããï¼ãepoll/kqueue/select ãï¼ã¤ä½¿ã£ã¦ã¤ãã³ãã«ã¼ããä½ãã ãã«ãã³ã¢CPUã§ã¹ã±ã¼ã«ããªãã®ã§ããµã¼ãã¼ã§ã¯ä»æãã®ã¢ãã«ã¯æµè¡ããªãã ã¯ã©ã¤ã¢ã³ãã§éåæãªã¡ãã»ã¼ã¸ã³ã°ãããããå ´åã¯ãã®ã¢ãã«ã使ããï¼ ãµã¼ãã¼ã«ã¡ãã»ã¼ã¸ãéä¿¡ ã¤ãã³ããã³ãã©ãç»é²ï¼ãã®ã¨ãã¤ãã³ããã³ãã©ã®ãã¤ã³ã¿ãåã£ã¦ãã ã¤ãã³ããã³ãã©->ãã©ã° ãONã«ãªãã¾ã§ã¤ãã³ãã«ã¼ããåã ã¤ãã³ããã³ãã©->çµæ ãè¿ã ï¼ã³ãã¯ã·ã§ã³ï¼ã¹ã¬ã
ã·ã³ãã«ã§ã¹ã±ã¼ã©ãã«ãªåæ£ã¹ãã¬ã¼ã¸ãèªç¤¾éçºããã©ã¤ãã㢠ä¸æ¹ã©ã¤ããã¢å·è¡å½¹CTOã®æ± éæºæ´æ°ã¯ï¼å社ã®ããã°ãåçæ稿ãµã¼ãã¹ãªã©ã®ã¤ã³ãã©ã§å©ç¨ä¸ã®ã¹ãã¬ã¼ã¸ä»®æ³åã½ãããèªç¤¾éçºããäºä¾ãç´¹ä»ãããã©ã¤ããã¢ã®ãµã¼ãã¹ç¾¤ãæ±ããè¦ä»¶ããããã«å®ä¾¡ã«å®¹éã追å ã§ããããéå°ãªæ©è½ã¨ä¿¡é ¼æ§ã¯ä¸è¦ãï¼æ± éæ°ï¼ã¨å¤æãã¡ã¼ã«ã¼è£½ã®ãããã¯ã¼ã¯ã»ã¹ãã¬ã¼ã¸ã®å©ç¨ãæ¢ãï¼ããã¡ã¤ã«ã®ãã¹ããã®ã¾ã¾URLã«ãªãããï¼ãã¡ã¤ã«ã»ã·ã¹ãã ã®ãã¹ãURLã«å¤æããªãã¦æ¸ããHTTPã§å ¥åºåããåæ£åä»®æ³ã¹ãã¬ã¼ã¸ã®éçºã«è¸ã¿åã£ãã®ã ã¨ããï¼åç4ï¼ã è¨è¨ææ³ã¯ãè¤æ°ãã¼ãéã®ä¸è²«æ§ã¯CAPå®çã«åºã¥ãã¦é 延ã妥åãï¼ã¹ã±ã¼ã©ããªãã£ã¨èªã¿åºãã®éãã«ãã ãã£ããä¸æ¹ã§æ¸ãè¾¼ã¿ã¯ããããã®é度ã§ããï¼èªè¨¼ã¨ã¢ã¯ã»ã¹å¶å¾¡ã¯ã¢ããªã±ã¼ã·ã§ã³ã§å®è£ ããã®ã§ä¸è¦ãï¼æ± éæ°ï¼ã¨ãããã®ãHTTPãµã¼ãã¼
2009/05/12 ç±³æ°è社大æã®ãã¥ã¼ã¨ã¼ã¯ã»ã¿ã¤ã ãºã¯5æ11æ¥ãRubyã«ãã大è¦æ¨¡åæ£å¦çã®ãã¼ã«ããããMap/Reduce Toolkitãï¼MRToolkitï¼ãGPLv3ã®ä¸ã«ãªã¼ãã³ã½ã¼ã¹ã§å ¬éããã¨çºè¡¨ãããMRToolkitã¯ããã§ã«ç¨¼åãã¦ããã¯ã©ã¹ã¿ä¸ã®Hadoopã¨åããã¦ä½¿ããã¨ã§Rubyã§å®¹æã«Map/Reduceå¦çãè¨è¿°ãããã¨ãã§ããä¸ç¨®ã®ã©ããã¼ãå¦çèªä½ã¯Hadoopãè¡ãããã§ã«Hadoopã使ã£ã¦ããã¦ã¼ã¶ã¼ã§ããã°ãä¸å°è¦æ¨¡ã®ããã¸ã§ã¯ãã«å¯¾ãã¦ãããã«MRToolkitãé©ç¨å¯è½ã¨ãã¦ããã ããã©ã«ãã§æç¨ãªMapãReduceã®å¦çã¢ã¸ã¥ã¼ã«ãå«ã¾ãã¦ãã¦ãæ°è¡ã®Rubyã¹ã¯ãªãããæ¸ãã ãã§ãä¾ãã°è¨å¤§ãªApacheã®ãã°ããIPã¢ãã¬ã¹å¥ã®é²è¦§å±¥æ´ãã¾ã¨ããã¨ãã£ãå¦çãå¯è½ã¨ãããç¬èªã«MapãReduceã®å¦çãå®ç¾©ãããã¨ã
ã¤ãã¼æ ªå¼ä¼ç¤¾ã¯ã2023å¹´10æ1æ¥ã«LINEã¤ãã¼æ ªå¼ä¼ç¤¾ã«ãªãã¾ãããLINEã¤ãã¼æ ªå¼ä¼ç¤¾ã®æ°ããããã°ã¯ãã¡ãã§ããLINEã¤ãã¼ Tech Blog ããã«ã¡ã¯ãªã¼ã¯ã·ã§ã³äºæ¥é¨ãã©ãããã¼ã æè¡ã®ãã£ãã¯ã§ãã ãªã¼ã¯ã·ã§ã³ã§ã¯ä¸é¨ãµã¼ãã¹ã« RDBMS ã® MySQL ã使ã£ã¦ãµã¼ãã¹ããæä¾ããã¦ããã ãã¦ããã¾ãã ãªã¼ã¯ã·ã§ã³ã§ã¯å¤ãã®ã客æ§ããã¢ã¯ã»ã¹ãé ãã¦ããã¾ãã®ã§ã大éã®æ´æ°ãåç §ã®å¦çé度ã«åªãã MySQL ãé¸æããã客æ§ã«ã¹ãã¬ã¹ãªããµã¼ãã¹ããå©ç¨ããã ãããã æ¥ã æ¥åã«åãçµã¾ããã¦ããã ãã¦ããã¾ãã ããããç²¾å¯æ©å¨ã«ã¯æ éãã¤ããã®ã§ããµã¼ãã¹éç¨ã®è¦³ç¹ãã㯠ãæ©å¨ãæ éããã®ã¯ããããªããããããããããã«æ©ã復æ§ããããã ã¨ãã£ããã¨ã念é ã«å ¥ãã¦ããã¾ãã å®éã«ã¯ãé害ãèµ·ãã£ã¦ãã復æ§ãããã®ã§ã¯ãªããé害ãçºçããå ´åã«
HowFriendFeedUsesMySqlToStoreSchemaLessData - FriendFeed ã§ã¯ MySQL ã使ãã©ã®ããã«ã¹ãã¼ãã¬ã¹ã®ãã¼ã¿ãä¿åãã¦ããã®ã ç®æ¬¡ ãã®è¨äºã«ã¤ã㦠FriendFeed? ã§ã¯ MySQL ã使ãã©ã®ããã«ã¹ãã¼ãã¬ã¹ã®ãã¼ã¿ãä¿åãã¦ããã®ã èæ¯ æ¦è¦³ 詳細 ä¸è²«æ§ã¨ååæ§ æ§è½ FriendFeed? ã§ã¯ MySQL ã使ãã©ã®ããã«ã¹ãã¼ãã¬ã¹ã®ãã¼ã¿ãä¿åãã¦ããã®ã ãã®è¨äºã«ã¤ã㦠"How FriendFeed? uses MySQL to store schema-less data" ã®æ¥æ¬èªè¨³ã§ã http://bret.appspot.com/entry/how-friendfeed-uses-mysql CC 2.5 ã§ã©ã¤ã»ã³ã¹ããã¦ãã¾ã: http://creativecommons.org/
æ¤ç´¢ããããã¬ã¼ãºãå ¥ããã°å³åº§ã«çµæãè¿ãã¦ããããã®Googleã§ããããã®1ãã¬ã¼ãºãå¦çãããããå®ã«1000å°ãã®ãµã¼ãã使ããããã0.2ç§ã§è¶ é«éå¦çãã¦ãããã¨ããWSDM 2009ã«ã¦æããã«ãªãã¾ãããåºèª¿è¬æ¼ãè¡ã£ãã®ã¯Googleãã§ãã¼ã§ããJeff Deanæ°ã§ã2008å¹´6æã«ããããGoogle I/Oãã«ã³ãã¡ã¬ã³ã¹ã§ã¯700ï½1000å°ã®ãµã¼ãã§0.5ç§ä»¥ä¸ã®æéããããã¨è¨ã£ã¦ãã¾ããããä»åã®è¬æ¼ã§ã¯ã¦ã¼ã¶ã¼ã®æ°ã¥ããªãã¨ããã§Googleã¯çå®ã«é²åãç¶ãã¦ãããã¨ãæããã«ãªãã¾ããã ç¥ããããGoogleã®è£å´ã®ææ°æ å ±ã¯ä»¥ä¸ããã Geeking with Greg: Jeff Dean keynote at WSDM 2009 Single Google Query uses 1000 Machines in 0.2 seconds ã¾ã
Twitterã®ã¹ã±ã¼ã«é¢ä¿ã§ãé¢ç½ãè¨äºãçºè¦ããã®ã§ã¾ã¨ãã ä¸ææãã¹ã±ã¼ã«ããªããã¨ããåä½ãä¸å®å®ãã ã¨ãè¨ããç¶ãã¦ããTwitterã5æããã«slashdot.jpã§ã話é¡ã«ãªã£ã¦ãããè«èª¿ã¯ç·ãã¦ãTwitterãã¹ã±ã¼ã«ããªãã®ã¯ãRubyã使ã£ã¦ãããããã¨ãããã®ã ã¨ãããåãï¼æããWhy Can't Twitter Scale? Blaine Cook Tries To Explainï¼ãªãã§Twitterã£ã¦ã¹ã±ã¼ã«ããªãã®ï¼ï¼ãã¨ãããblogç´¹ä»è¨äºãSilicon Alley Insiderã«æ²è¼ããããè¨äºã®å ã«ãªã£ãblogã¨ã³ããªã¯ãTwitterã®åãã¼ãã¢ã¼ããã¯ãã ã£ãBlaine Cookæ°ã«ãããã®ãCookæ°ã«ããã°ãTwitterã®ã¹ã±ã¼ã«ã¨Rubyã¯ä½ã®é¢ä¿ããªãã¨ããã Why Can't Twitter Scale? Blai
ãªãªã¼ã¹ãé害æ å ±ãªã©ã®ãµã¼ãã¹ã®ãç¥ãã
ææ°ã®äººæ°ã¨ã³ããªã¼ã®é ä¿¡
å¦çãå®è¡ä¸ã§ã
j次ã®ããã¯ãã¼ã¯
kåã®ããã¯ãã¼ã¯
lãã¨ã§èªã
eã³ã¡ã³ãä¸è¦§ãéã
oãã¼ã¸ãéã
{{#tags}}- {{label}}
{{/tags}}