[B! åˆ†æ•£å‡¦ç†] terurouã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

terurou id:terurou

åˆ†æ•£å‡¦ç†ã«é–¢ã™ã‚‹terurouã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (10)

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

${{author_name}}$
{{author_name}}{{created}}
{{ #comment }}{{ comment }}{{ /comment }}
- {{ label }}

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

Raft Consensus Algorithm
What is Raft? Raft is a consensus algorithm that is designed to be easy to understand. It's equivalent to Paxos in fault-tolerance and performance. The difference is that it's decomposed into relatively independent subprobl ems, and it cleanly addresses all major pieces needed for practical systems. We hope Raft will make consensus available to a wider audience, and that this wider audience will be
terurou 2019/04/29
raft

åˆ†æ•£å‡¦ç†
ãƒªãƒ³ã‚¯
ã‚ªãƒ³ãƒ©ã‚¤ãƒ³ã‚²ãƒ¼ãƒ ã®ä»•çµ„ã¿ã¨å·¥å¤«
ã‚ªãƒ³ãƒ©ã‚¤ãƒ³ã‚²ãƒ¼ãƒ ã®ä»•çµ„ã¿ã‚„å·¥å¤«ã‚’èª¿ã¹ã¦ã¿ãŸã®ã‚’ç¤¾å†…å‹‰å¼·ä¼šã§ç™ºè¡¨ã—ãŸã€‚ã¨ãã®ã‚¹ãƒ©ã‚¤ãƒ‰ã€‚ã®å…¬é–‹ç”¨ã€‚ ã‚ªãƒ³ãƒ©ã‚¤ãƒ³ã‚²ãƒ¼ãƒ ã®ç¨®åˆ¥ã¨ãã‚Œãžã‚Œã®ä»•çµ„ã¿ã«ã¤ã„ã¦ã®è©±ã¨ã€ã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹ã«ãªã£ã¦ã„ã‚‹Quakeã®ä»•çµ„ã¿ã®è©±ã€ã¨ã„ã†2ã¤ã®è©±ãŒä¸»ãªãƒˆãƒ”ãƒƒã‚¯
terurou 2015/07/13
game

messaging

åˆ†æ•£å‡¦ç†
ãƒªãƒ³ã‚¯
åˆ†æ•£ã‚·ã‚¹ãƒ†ãƒ å‡¦ç†ãƒ¢ãƒ‡ãƒ«ã«é–¢ã™ã‚‹å‹•å‘ã«ã¤ã„ã¦ï¼ˆMapReduceã‹ã‚‰Borgã¾ã§ï¼‰
è©³ç´°ã«ã¤ã„ã¦ã¯å¾Œè¿°ã—ã¾ã™ãŒã€MapReduceã®å‡¦ç†ãƒ¢ãƒ‡ãƒ«ã¯ã€ä¸Šè¨˜ã®é€šã‚Šå„åŒºåˆ†ã”ã¨ã«ãã‚Œãžã‚Œå˜ç´”åŒ–ï¼ˆé™å®šï¼‰ã•ã‚ŒãŸãƒ¢ãƒ‡ãƒ«ã§ã‚ã£ãŸã¨è¨€ãˆã¾ã™ã€‚ ã¾ãŸã€MapReduceã®é–¢æ•°ãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ãŠã‚ˆã³ã‚°ãƒ©ãƒ•çš„ãªç‰¹å¾´ã‚‚åˆã‚ã›ã¦ä»¥ä¸‹ã«æ•´ç†ã—ã¦ã¿ã¾ã™ã€‚ é–¢æ•°ãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°çš„ãªç‰¹å¾´ MapãŠã‚ˆã³Reduceãƒ•ã‚§ãƒ¼ã‚ºã¯ã€ãã‚Œãžã‚Œé–¢æ•°åž‹ãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ã®MapãŠã‚ˆã³Reduceå‡¦ç†ã‚’ãƒ¢ãƒ‡ãƒ«åŒ–ã—ãŸã‚‚ã®ã§ã™ã€‚MapReduceã¯ã€å‚ç…§é€éŽæ€§ãŒã‚ã‚‹ç´”ç²‹ãªé–¢æ•°å‡¦ç†ã¨è¨€ãˆã¾ã™ã€‚å‚ç…§é€éŽæ€§ã¨ã¯å…¥åŠ›ã«ã‚ˆã‚Šå‡ºåŠ›ãŒä¸€æ„ã«æ±ºã¾ã‚‹æ€§è³ªã®ã“ã¨ã§ã™ã€‚è¨€ã„æ›ãˆã‚Œã°MapReduceã®å‡¦ç†ã¯ã€å¤§åŸŸãªã©ã®å‡¦ç†ã«å½±éŸ¿ã™ã‚‹å¤–éƒ¨ã®ç’°å¢ƒã¯æŒãŸãšã€å†…éƒ¨çš„ã«ã‚‚é™çš„ãªä¸€æ™‚å¤‰æ•°ãªã©ã®çŠ¶æ…‹ã‚‚æŒãŸãªã„ã“ã¨ã‚’æ„å‘³ã—ã¾ã™ã€‚ ç´”ç²‹ãªé–¢æ•°å‡¦ç†ã¯è¤‡æ•°ã®å‡¦ç†ãŒåŒæ™‚ã«å®Ÿè¡Œã•ã‚Œã¦ã‚‚ä»–ã®ä¸¦åˆ—ã«å‹•ä½œã—ã¦ã„ã‚‹å‡¦ç†ã®çŠ¶æ…‹ã«ã¯å·¦å³ã•ã‚Œãªã„ãŸã‚ã€ã“ã®å‚ç…§é€éŽæ€§ã¯ä¸¦åˆ—åŒ–ã«å‘ã„ã¦ã„ã‚‹æ€§è³ªãŒã‚ã‚Šã¾
terurou 2015/06/11
hadoop

åˆ†æ•£å‡¦ç†
ãƒªãƒ³ã‚¯
åˆ†æ•£ã‚·ã‚¹ãƒ†ãƒ ã®ä¸€è²«æ€§ã«é–¢ã™ã‚‹å‹•å‘ã«ã¤ã„ã¦
ãƒ¤ãƒ•ãƒ¼æ ªå¼ä¼šç¤¾ã¯ã€2023å¹´10æœˆ1æ—¥ã«LINEãƒ¤ãƒ•ãƒ¼æ ªå¼ä¼šç¤¾ã«ãªã‚Šã¾ã—ãŸã€‚LINEãƒ¤ãƒ•ãƒ¼æ ªå¼ä¼šç¤¾ã®æ–°ã—ã„ãƒ–ãƒã‚°ã¯ã“ã¡ã‚‰ã§ã™ã€‚LINEãƒ¤ãƒ•ãƒ¼ Tech Blog ã‚·ã‚¹ãƒ†ãƒ çµ±æ‹¬æœ¬éƒ¨ã‚¢ãƒ¼ã‚ãƒ†ã‚¯ãƒˆå®¤ ä»Šé‡Žã§ã™ã€‚ æ˜¨å¹´ã¯ã€Twitter,Facebookã‚’å§‹ã‚ã¨ã™ã‚‹ã‚¯ãƒ©ã‚¦ãƒ‰å„ç¤¾ã§æ–°è¦ã®åˆ†æ•£ã‚·ã‚¹ãƒ†ãƒ é–‹ç™ºã®ãƒ—ãƒã‚¸ã‚§ã‚¯ãƒˆãŒç›¸æ¬¡ã„ã§ç™ºè¡¨ã•ã‚ŒãŸå¹´ã§ã—ãŸã€‚ã“ã‚Œã‚‰ã®æ–°ã—ã„åˆ†æ•£ã‚·ã‚¹ãƒ†ãƒ ã‚’é–‹ç™ºã™ã‚‹ç†ç”±ã‚„ã€ãã®èƒŒæ™¯ã«ã‚ã‚‹ã‚‚ã®ã¯ä½•ãªã®ã§ã—ã‚‡ã†ã‹ï¼Ÿ ä»Šå›žã¯ã€æ˜¨å¹´æœ«ã«é–‹å‚¬ã•ã‚ŒãŸé«˜ä¿¡é ¼æ€§åˆ†æ•£ã‚·ã‚¹ãƒ†ãƒ ç³»ã®å›½éš›å¦ä¼šã§ã‚ã‚‹SRDS 2014[1]ã®ç™ºè¡¨å†…å®¹ã«é–¢é€£ã™ã‚‹è«–æ–‡ã®è©±é¡Œã‚‚è¸ã¾ãˆã¦ã€æ˜¨ä»Šã®ã‚¯ãƒ©ã‚¦ãƒ‰å„ç¤¾ã®åˆ†æ•£ã‚·ã‚¹ãƒ†ãƒ ã®å‹•å‘ã«ã¤ã„ã¦æ•´ç†ã—ã¦ã¿ã¾ã™ã€‚ åˆ†æ•£ã‚·ã‚¹ãƒ†ãƒ ã«ãŠã‘ã‚‹ã‚¯ãƒ©ã‚¦ãƒ‰å„ç¤¾ã®å‹•å‘ è¿‘å¹´ã®åˆ†æ•£ãƒ‡ãƒ¼ã‚¿ãƒ™ãƒ¼ã‚¹ã®ä¸–ç•Œã§ã¯ã€Amazonã®Dynamo[2]ã‚„Facebookã®Cassandra[3]ãªã©ã‚’ä»£è¡¨ã¨ã™ã‚‹çµæžœæ•´åˆæ€§ï¼ˆEve
terurou 2015/04/06
æœ€è¿‘ã®å‹•å‘ã‚’ã¡ã‚ƒã‚“ã¨è¿½ãˆã¦ã„ãªã„éƒ¨åˆ†ã ã£ãŸã®ã§éžå¸¸ã«åŠ©ã‹ã‚‹

åˆ†æ•£å‡¦ç†

database

transaction

CAPå®šç†

atmic
ãƒªãƒ³ã‚¯
MapReduceä»¥å¤–ã®åˆ†æ•£å‡¦ç†åŸºç›¤BSP, Piccolo, Sparkã®ç´¹ä»‹ - Preferred Networks Tech Blog
ã©ã†ã‚‚ï¼Œå®Ÿã¯ä»Šå¹´ã‹ã‚‰é–‹ç™ºãƒãƒ¼ãƒ ã«joinã—ã¦ã„ãŸä¸å·ã§ã™ï¼Žå¯æ„›ã„çŠ¬ã®å†™çœŸãŒãªã‹ã£ãŸã®ã§ï¼Œå¯æ„›ã„ãƒžã‚¹ã‚³ãƒƒãƒˆã®ç”»åƒã‚’è²¼ã£ã¦ãŠãã¾ã™ï¼Ž æœ€è¿‘MapReduceã¨ã‹ãã®å®Ÿè£…ã§ã‚ã‚‹Hadoopã¨ã‹ã‚’ã‚ˆãèžãã‚ˆã†ã«ãªã‚Šã¾ã—ãŸï¼Žã“ã‚Œã¯ã¤ã¾ã‚Šï¼Œãã‚Œã ã‘å¤§é‡ã®ãƒ‡ãƒ¼ã‚¿ã‚’ãªã‚“ã¨ã‹å‡¦ç†ã—ãŸã„ã¨ã„ã†è¦æœ›ãŒã‚ã‚‹ã‹ã‚‰ã ã¨æ€ã„ã¾ã™ï¼Žã—ã‹ã—å½“ãŸã‚Šå‰ã§ã™ãŒï¼ŒMapReduceã¯éŠ€ã®å¼¾ä¸¸ã§ã¯ã‚ã‚Šã¾ã›ã‚“ï¼Ž ã¨ã„ã†ã“ã¨ã§ï¼Œæœ€è¿‘æ°—ã«ãªã£ã¦ã„ã‚‹MapReduceã¨ã¯é•ã£ãŸã‚¢ãƒ—ãƒãƒ¼ãƒã‚’å–ã£ã¦ã„ã‚‹åˆ†æ•£å‡¦ç†åŸºç›¤ã«ã¤ã„ã¦ï¼Œç¤¾å†…ã®TechTalkã§è©±ã—ãŸå†…å®¹ã‚’ç°¡å˜ã«ã¾ã¨ã‚ã¦ç´¹ä»‹ã—ãŸã„ã¨æ€ã„ã¾ã™ï¼Ž Bulk Sychronous Parallel ã“ã®ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ è‡ªä½“ã¯1990å¹´ã«èª•ç”Ÿã—ãŸã‚‚ã®ã§ã™ï¼Žé•·ã„ã®ã§BSPã¨æ›¸ãã¾ã™ï¼Žã•ã¦ï¼Œã‚°ãƒ©ãƒ•ã‹ã‚‰æœ€çŸçµŒè·¯ã‚’æ±‚ã‚ã‚‹æ™‚ï¼ŒMapReduceã¯ä½¿ãˆã‚‹ã§ã—ã‚‡ã†ã‹ï¼Ÿã“ã®ã‚ˆã†ãªè«–æ–‡ãŒå‡ºã‚‹ãã‚‰ã„ã§ã™ã‹ã‚‰å‡ºæ¥ãªã„ã“ã¨ã¯ã‚
terurou 2011/06/17
åˆ†æ•£å‡¦ç†

BSP

Piccolo

Spark
ãƒªãƒ³ã‚¯
æœ¬ã‚’èªã‚€ GREE Labsã§Hadoopã®è©±ã‚’èžã„ã¦ããŸ
GREEã•ã‚“ã§ä¸å®šæœŸã§ã‚„ã£ã¦ã‚‹ã€GREE Labsã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹ãƒ†ã‚¯ãƒŽãƒã‚¸ãƒ¼å‹‰å¼·ä¼šã§ã€Hadoopã®è©±ã‚’èžã„ã¦ãã¾ã—ãŸã€‚Hadoopã¯ã€ã¤ã¾ã‚Šã¯Googleã®GFSã‚„MapReduceã®ã‚¯ãƒãƒ¼ãƒ³ã ãã†ã§ã€ã€ŒGoogleã‚’æ”¯ãˆã‚‹æŠ€è¡“ã€ã«ãƒˆã‚ãƒ¡ã„ãŸäººãªã‚‰å¿…è¦‹ã§ã™ãã€‚ ç™ºè¡¨ã¯ã€æŠ€è¡“é¢ã‚’ç°¡æ½”ã«æŠ¼ã•ãˆãŸã†ãˆã§ã‚ã‹ã‚Šã‚„ã™ãã€ãã®ã†ãˆå®Ÿéš›ã®åˆ©ç”¨äº‹ä¾‹ã®è©±ã‚‚èžã‘ãŸã®ãŒé¢ç™½ã‹ã£ãŸã¨æ€ã„ã¾ã™ã€‚æœ€è¿‘ã®Webç³»ã§ã¯ã€ã‚µãƒ¼ãƒ“ã‚¹é¢ã§ã‚‚ãƒžãƒã‚¿ã‚¤ã‚ºé¢ã§ã‚‚ã€ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°ã¨ã‹è¡Œå‹•ã‚¿ãƒ¼ã‚²ãƒ†ã‚£ãƒ³ã‚°ã¨ã‹ãŒã‚¢ãƒ„ã„ã‚“ã ãªãã¨æ€ã„ã¾ã—ãŸã€‚ ãƒ—ãƒ¬ã‚¼ãƒ³è³‡æ–™ã‚‚ustéŒ²ç”»ã‚‚å…¬é–‹ã•ã‚Œã¦ã„ã¾ã™ãŒã€ä»¥ä¸‹ã€è‡ªåˆ†ã®ãƒ¡ãƒ¢ã¨ã„ã†æ„å‘³ã§è¨˜éŒ²ã—ã¦ãŠãã¾ã™ã€‚ Hadoopã«ã¤ã„ã¦ï¼ˆå¤ªç”°ä¸€æ¨¹ï¼‰ Preferred Infrastructureã®CTOã§ã€Sedueã®ä½œè€…ã€‚å¤§é‡ã®ãƒ‡ãƒ¼ã‚¿ã®å‡¦ç†ãŒãƒ†ãƒ¼ãƒžã§ã€åŠåˆ†ã¯é…’ã§ã§ãã¦ã„ã‚‹wã€‚ãã‚“ãªã“ã‚“ãªã§ã€ã¯ã¦ãƒ–æ¤œç´¢ã§ã‚‚ä½¿ã‚ã‚Œã¦ã„
terurou 2008/11/28
hadoop

java

EC2

event

åˆ†æ•£å‡¦ç†
ãƒªãƒ³ã‚¯
ä¸¸å±±ä¸äºŒå¤«ãƒ¬ã‚¯ãƒãƒ£ãƒ¼ã‚·ãƒªãƒ¼ã‚ºç¬¬1å›ž
æ—¥æœ¬Javaãƒ¦ãƒ¼ã‚¶ãƒ¼ã‚°ãƒ«ãƒ¼ãƒ—ã‚„æ—¥æœ¬Androidã®ä¼šã®ä¼šé•·ã‚’å‹™ã‚ã€ã‚¯ãƒ©ã‚¦ãƒ‰ã«é–¢ã™ã‚‹å‹‰å¼·ä¼šãªã©ã‚‚ç²¾åŠ›çš„ã«é–‹å‚¬ã•ã‚Œã¦ã„ã‚‹ä¸¸å±±ä¸äºŒå¤«å…ˆç”Ÿï¼ˆæ—©ç¨²ç”°å¤§å¦å¤§å¦é™¢æƒ…å ±ç”Ÿç”£ã‚·ã‚¹ãƒ†ãƒ ç ”ç©¶ç§‘å®¢å“¡æ•™æŽˆï¼‰ã€‚ãã‚“ãªä¸¸å±±å…ˆç”ŸãŒç¤¾ä¼šäººã®IT æŠ€è¡“è€…ã‚’å¯¾è±¡ã«ã€å¤šå½©ãªã‚²ã‚¹ãƒˆã‚¹ãƒ”ãƒ¼ã‚«ãƒ¼ã‚’è¿Žãˆã€æœ€å…ˆç«¯ã®æŠ€è¡“å‹•å‘ã‚’å–ã‚Šä¸Šã’ã‚‹ã€Œä¸¸å±±å…ˆç”Ÿãƒ¬ã‚¯ãƒãƒ£ãƒ¼ã‚·ãƒªãƒ¼ã‚º in æ±äº¬ã€ã‚’ä»Šå¹´ã‚‚é–‹å‚¬è‡´ã—ã¾ã™ã€‚ ä»Šå¹´åº¦ã§7å¹´ç›®ã‚’è¿Žãˆã‚‹åŒã‚»ãƒŸãƒŠãƒ¼ã€‚å›žã«ã‚ˆã£ã¦ã¯æŠ½é¸ã¨ãªã‚‹ã“ã¨ã‚‚çã—ããªã„ã»ã©ã€æœ‰ç”¨ãªã‚»ãƒŸãƒŠãƒ¼ã¨ã—ã¦çŸ¥ã‚‰ã‚Œã¦ã„ã¾ã™ã€‚ä»Šå¹´åº¦ã‚‚ã€ã‚¨ãƒ³ã‚¿ãƒ¼ãƒ—ãƒ©ã‚¤ã‚ºåˆ†é‡Žã®å„ªã‚ŒãŸè«–å®¢ã«æ•°å¤šãç™»å£‡ã„ãŸã ãäºˆå®šã§ã€èˆˆå‘³æ·±ã„ãƒ¬ã‚¯ãƒãƒ£ãƒ¼ã¨ãªã‚‹ã“ã¨ã§ã—ã‚‡ã†ã€‚ãœã²ã€ç”³ã—è¾¼ã¿ã®ä¸Šã€ä¼šå ´ã¸ã¨è¶³ã‚’ãŠé‹ã³ãã ã•ã„ã€‚ ä»¥ä¸‹ã€ä¸¸å±±å…ˆç”Ÿã‹ã‚‰ã®ãƒ¡ãƒƒã‚»ãƒ¼ã‚¸ã‚’æŠœç²‹ã—ã¾ã™ã€‚ ãƒ†ãƒ¼ãƒžï¼šã€Œã‚¯ãƒ©ã‚¦ãƒ‰ã€SOAã€ãƒ¢ãƒã‚¤ãƒ«ãƒ‡ãƒã‚¤ã‚¹ã€ ä»Šå¹´ã®ä¸¸ãƒ¬ã‚¯ã‚»ãƒŸãƒŠãƒ¼ã¯ã€ã‚¯ãƒ©ã‚¦ãƒ‰ã‚’ãƒ¡ã‚¤ãƒ³ã®ãƒ†ãƒ¼ãƒžã«ã—ãŸã„ã¨æ€ã„ã¾ã™ã€‚ ã‚¯ãƒ©ã‚¦ãƒ‰ã‚³ãƒ³ãƒ”
terurou 2008/11/06
ã™ã”ãå‚åŠ ã—ãŸã„ã€‚

Cloud

SaaS

SOA

åˆ†æ•£å‡¦ç†

event
ãƒªãƒ³ã‚¯
Home | Hypertable - Big Data. Big Performance
Big Data. Big Performance. Hyperta ble delivers scala ble database capacity at maximum performanceÂ to speed up your big data application and reduce your hardware footprint. 100% Hadoop Compatible Hyperta ble seamlessly overlays on top of Hadoop to provide superchargedÂ scala ble database infastructure for your big data application. Less Hardware, Lower Cost Hyperta ble delivers maximum efficiency and su
terurou 2008/02/11
database

BigTable

åˆ†æ•£å‡¦ç†

Hypertable
ãƒªãƒ³ã‚¯
ConsistentHashing - ã‚³ãƒ³ã‚·ã‚¹ãƒ†ãƒ³ãƒˆãƒ»ãƒãƒƒã‚·ãƒ¥æ³•
ConsistentHashing - ã‚³ãƒ³ã‚·ã‚¹ãƒ†ãƒ³ãƒˆãƒ»ãƒãƒƒã‚·ãƒ¥æ³• ç›®æ¬¡ ã“ã®æ–‡æ›¸ã«ã¤ã„ã¦ ã‚³ãƒ³ã‚·ã‚¹ãƒ†ãƒ³ãƒˆãƒ»ãƒãƒƒã‚·ãƒ¥æ³• å®Ÿä¾‹ å®Ÿè£… ç”¨é€” ã‚³ãƒ³ã‚·ã‚¹ãƒ†ãƒ³ãƒˆãƒ»ãƒãƒƒã‚·ãƒ¥æ³• ã“ã®æ–‡æ›¸ã«ã¤ã„ã¦ "Tom White's Blog: Consistent Hashing" ã®æ—¥æœ¬èªžè¨³ã§ã™. http://weblogs.java.net/blog/tomwhite/archive/2007/11/consistent_hash.html æŽ¨æ•²æ“è¿Ž: èª¤è¨³, ã‚¿ã‚¤ãƒ, è¨³èªžã®ä¸çµ±ä¸€, ãã®ã»ã‹... åŽŸæ–‡ã®ãƒ©ã‚¤ã‚»ãƒ³ã‚¹: http://creativecommons.org/licenses/by-nc-sa/2.0/ ç§ã¯ä»Šã¾ã§ã«ä½•åº¦ã‹ã‚³ãƒ³ã‚·ã‚¹ãƒ†ãƒ³ãƒˆãƒ»ãƒãƒƒã‚·ãƒ¥æ³•ã«ã¨ã‚Šãã‚“ã ã“ã¨ãŒã‚ã‚‹ã€‚ ã“ã®ã‚¢ã‚¤ãƒ‡ã‚¢ã‚’ã‚ã‚‰ã‚ã—ãŸè«–æ–‡ ( David Karger ã‚‰ã«ã‚ˆã‚‹ Consistent Hashing and R
terurou 2007/12/28
ã‚„ã£ã¨èªã‚“ã ã€‚ä»Šåº¦ã¤ã‹ã†ã‹ã‚‚ã€‚

ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ

cache

hash

scalability

åˆ†æ•£å‡¦ç†

ConsistentHashing
ãƒªãƒ³ã‚¯
å¤§è¦æ¨¡åˆ†æ•£å‡¦ç†å‘ã‘ã®å›½ç”£â€œã‚¦ã‚§ãƒ–OSâ€ã‚’Rubyã§é–‹ç™ºä¸ âˆ’ ï¼ IT
2007/11/26 2007å¹´11æœˆ24æ—¥ã€ã€Œæ¥½å¤©ãƒ†ã‚¯ãƒŽãƒã‚¸ãƒ¼ã‚«ãƒ³ãƒ•ã‚¡ãƒ¬ãƒ³ã‚¹2007ã€ã«ãŠã„ã¦ã€Rubyè¨€èªžã®é–‹ç™ºè€…ã§æ¥½å¤©æŠ€è¡“ç ”ç©¶æ‰€ãƒ•ã‚§ãƒãƒ¼ã®ã¾ã¤ã‚‚ã¨ã‚†ãã²ã‚æ°ã¯ã€é–‹ç™ºä¸ã®å¤§è¦æ¨¡åˆ†æ•£å‡¦ç†åŸºç›¤ã€ŒRomaã€ï¼ˆãƒãƒ¼ãƒžï¼‰ã¨ã€ŒFairyã€ï¼ˆãƒ•ã‚§ã‚¢ãƒªãƒ¼ï¼‰ã®ã‚³ãƒ³ã‚»ãƒ—ãƒˆã‚’èªžã£ãŸã€‚ç ”ç©¶æ®µéšŽã§ã¯ã‚ã‚‹ã‚‚ã®ã®ã€ç±³ã‚°ãƒ¼ã‚°ãƒ«ãªã©ä¸–ç•Œãƒˆãƒƒãƒ—ã‚¯ãƒ©ã‚¹ã®ãƒãƒƒãƒˆä¼æ¥ã ã‘ãŒæŒã¤å¤§è¦æ¨¡åˆ†æ•£å‡¦ç†æŠ€è¡“ã«çœŸã£å‘ã‹ã‚‰æŒ‘æˆ¦ã™ã‚‹è©¦ã¿ã ã€‚ ç±³ã‚°ãƒ¼ã‚°ãƒ«ã€ç±³ãƒ¤ãƒ•ãƒ¼ã€ç±³ã‚¢ãƒžã‚¾ãƒ³ãªã©ä¸–ç•Œãƒˆãƒƒãƒ—ã‚¯ãƒ©ã‚¹ã®å¤§æ‰‹ãƒãƒƒãƒˆä¼æ¥ã¯ã€å·¨å¤§ãªãƒˆãƒ©ãƒ•ã‚£ãƒƒã‚¯ã«å¯¾å‡¦ã™ã‚‹ãŸã‚ã€å¤§è¦æ¨¡ãƒ‡ãƒ¼ã‚¿ã‚»ãƒ³ã‚¿ãƒ¼ã®ä¿¡é ¼æ€§ã€å¯ç”¨æ€§ã€æ€§èƒ½ãªã©ã‚’ç¢ºä¿ã™ã‚‹å¤§è¦æ¨¡åˆ†æ•£å‡¦ç†åŸºç›¤ã®ç ”ç©¶é–‹ç™ºã‚’é€²ã‚ã¦ã„ã‚‹ã€‚æœ€è¿‘ã§ã¯ã€ã“ã†ã—ãŸå¤§è¦æ¨¡åˆ†æ•£å‡¦ç†åŸºç›¤ã¯â€œã‚¦ã‚§ãƒ–OSâ€ã¨å‘¼ã°ã‚Œã‚‹ã“ã¨ã‚‚ã‚ã‚Šã€æ³¨ç›®ã‚’é›†ã‚ã¦ã„ã‚‹ã€‚ã¤ã¾ã‚Šä¸–ç•Œãƒˆãƒƒãƒ—ã‚¯ãƒ©ã‚¹ã®ãƒãƒƒãƒˆä¼æ¥ã¯â€œã‚¦ã‚§ãƒ–OSâ€ã‚’è‡ªç¤¾é–‹ç™ºã—ã¦ã„ã‚‹ã‚ã‘ã ãŒã€æ¥½å¤©ã§ã‚‚å›½ç”£ã®â€œã‚¦ã‚§ãƒ–OSâ€ãŒç”Ÿã¾ã‚Œ
terurou 2007/11/27
æ¥½å¤©ãŒRubyã§GFSã‚„ã‚‰Hadoopã¿ãŸã„ãªåˆ†æ•£ç’°å¢ƒã‚’é–‹ç™ºã™ã‚‹ã¨ã€‚ä»Šå¾Œã€Rubyè‡ªä½“æ‹¡å¼µã—ã¦ã‚’erlangã¿ãŸã„ãªåˆ†æ•£å‡¦ç†ãŒã§ãã‚‹ã‚ˆã†ã«ã¨ã‹ã™ã‚“ã®ã‹ãªã€‚/ã¨ã“ã‚ã§ã‚¦ã‚§ãƒ–OSã¨ã„ã†è¨€ã„æ–¹ã¯ãŠã‹ã—ã„ã ã‚ã€‚

ruby

event

report

æ¥½å¤©

Hadoop

GFS

åˆ†æ•£å‡¦ç†
ãƒªãƒ³ã‚¯
1