[B! bigtable] hirose31ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

hirose31 id:hirose31

bigtableã«é–¢ã™ã‚‹hirose31ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (9)

${{author_name}}$

{{{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}}

moratorium | PNUTS
PNUTS Yahooã®åºƒåŸŸåˆ†æ•£DBï¼Ÿã¾ã è«–æ–‡èªã‚“ã§ãªã„ã®ã§é–“é•ã£ã¦ãŸã‚‰ã™ã„ã¾ã›ã‚“ã€‚ PNUTS - Platform for Nim ble Universal Table Storage PNUTS: Yahoo!â€™s Hosted Data Serving Platform Data Challenges at Yahoo! DBã¯é–€å¤–æ¼¢ã ã‘ã©ã€VLDBã¨SIGMODãã‚‰ã„ã¯ç›®ã‚’é€šã—ã¦ãŠã‹ãªã„ã¨ãªã‚ã¨æ¿€ã—ãåçœã€‚ SIGMOD 2008ã§ã¯ã€Facebookã®åˆ†æ•£ã‚¹ãƒˆãƒ¬ãƒ¼ã‚¸Cassandraã€Googleã®MegaStore (BigTableã®ä¸Šã«ä½œã‚‰ã‚ŒãŸTransactionManager)ãªã‚“ã‹ã‚‚ç™ºè¡¨ãŒæœ‰ã£ãŸæ¨¡æ§˜ã€‚ GoogleAppEngineãŒBigTableã‚’ä½¿ã£ã¦ã‚‹ã®ã«TransactionãŒæœ‰ã£ãŸã®ã§ã€Œã‚“ï¼Ÿã€ã¨æ€ã£ãŸã‚“ã ã‘ã©ã€ã“ã†ã„ã†ãƒ¬ã‚¤ãƒ¤ãƒ¼ã‚’è¢«ã›ã¦ã„ã‚‹ã®ã§ã™ãã€‚ ãŸã—
hirose31 2008/09/17
PNUTS

bigtable

Cassandra

MegaStore
ãƒªãƒ³ã‚¯
Hadoopã€hBaseã§æ§‹ç¯‰ã™ã‚‹å¤§è¦æ¨¡åˆ†æ•£ãƒ‡ãƒ¼ã‚¿å‡¦ç†ã‚·ã‚¹ãƒ†ãƒ ï¼šCodeZine
ã¯ã˜ã‚ã«ã€€ã“ã®é€£è¼‰ã§ã¯ã€å¤§è¦æ¨¡åˆ†æ•£è¨ˆç®—ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã€ŒHadoopã€ã¨ã€ãã®ä¸Šã«ã¤ãã‚‰ã‚ŒãŸå¤§è¦æ¨¡åˆ†æ•£ãƒ‡ãƒ¼ã‚¿ãƒ™ãƒ¼ã‚¹ã€ŒhBaseã€ã®ä»•çµ„ã¿ã¨ç°¡å˜ãªã‚µãƒ³ãƒ—ãƒ«ã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ã‚’ç´¹ä»‹ã—ã¾ã™ã€‚Hadoopã¨hBaseã¯ã€Googleã®åŸºç›¤ã‚½ãƒ•ãƒˆã‚¦ã‚§ã‚¢ã®ã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹ã‚¯ãƒãƒ¼ãƒ³ã§ã™ã€‚æ©Ÿèƒ½ã‚„ã‚³ãƒ³ã‚»ãƒ—ãƒˆã«ã¤ã„ã¦ã¯ã€GoogleãŒç™ºè¡¨ã—ã¦ã„ã‚‹å¦è¡“è«–æ–‡ã«ä¾ã£ã¦ã„ã¾ã™ã€‚ã€€ã“ã‚Œã‚‰ã®å¦è¡“è«–æ–‡ã«ã‚ˆã‚‹ã¨ã€Googleã§ã¯å¤§è¦æ¨¡åˆ†æ•£ãƒ•ã‚¡ã‚¤ãƒ«ã‚·ã‚¹ãƒ†ãƒ ã€ŒGoogle File Systemã€ã€å¤§è¦æ¨¡åˆ†æ•£è¨ˆç®—ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã€ŒMapReduceã€ã€å¤§è¦æ¨¡åˆ†æ•£ãƒ‡ãƒ¼ã‚¿ãƒ™ãƒ¼ã‚¹ã€ŒBigTableã€ã€åˆ†æ•£ãƒãƒƒã‚¯ã‚µãƒ¼ãƒ“ã‚¹ã€ŒChubbyã€ã¨ã„ã†4ã¤ã®ã‚¤ãƒ³ãƒ•ãƒ©ã‚½ãƒ•ãƒˆã‚¦ã‚§ã‚¢ãŒä½¿ã‚ã‚Œã¦ã„ã¾ã™ã€‚ å›³1ã«Googleã®åŸºç›¤æŠ€è¡“é–“ã®ä¾å˜é–¢ä¿‚ã€ãã—ã¦ãã‚Œã«å¯¾å¿œã™ã‚‹OSSã®å¯¾å¿œé–¢ä¿‚ã‚’ç¤ºã—ã¾ã—ãŸã€‚ã¾ãšã¯å¯¾å¿œã™ã‚‹Googleã®åŸºç›¤æŠ€è¡“ãã‚Œãžã‚Œã®æ©Ÿèƒ½ã‚„ç‰¹å¾´ã‚’ã–ã£ãã‚Šã¨
hirose31 2008/05/01
hadoop

hBase

bigtable

mapreduce
ãƒªãƒ³ã‚¯
ãŸã‘ã¾ã‚‹ / Google App Engine ã®ãƒ‡ãƒ¼ã‚¿ã‚¹ãƒˆã‚¢ã¯ Bigtable ã‚’ã©ã®ã‚ˆã†ã«ä½¿ã£ã¦ã„ã‚‹ã®ã‹
_ Google App Engine ã®ãƒ‡ãƒ¼ã‚¿ã‚¹ãƒˆã‚¢ã¯ Bigtable ã‚’ã©ã®ã‚ˆã†ã«ä½¿ã£ã¦ã„ã‚‹ã®ã‹ [gae][bigtable] Google App Engine (GAE) ãŒç™ºè¡¨ã•ã‚Œã¦ã‹ã‚‰2é€±é–“ã»ã©çµŒã¡ã¾ã™ï¼ŽGFS ã‚„ Bigtable ã¨ã„ã†åå‰ã ã‘ã¯ã‚ˆãè€³ã«ã™ã‚‹ã‚ˆã†ã«ãªã‚Šã¾ã—ãŸãŒï¼ŒBigtable ã¨ GAE ã®ã‚®ãƒ£ãƒƒãƒ—ã«ã¤ã„ã¦ã¯è©±é¡Œã«ãªã£ã¦ã„ãªã„ã‚ˆã†ã«æ€ã„ã¾ã™ï¼Ž Bigtable ã¯ multi dimensional sorted table ã¨è¨€ã‚ã‚Œã‚‹ã‚ˆã†ã«ï¼Œ primary key (row key) ã§ã‚½ãƒ¼ãƒˆã•ã‚ŒãŸãƒ†ãƒ¼ãƒ–ãƒ«ã§ã—ã‹ã‚ã‚Šã¾ã›ã‚“ï¼Žã¤ã¾ ã‚Šï¼ŒGAE ã®ãƒ‡ãƒ¼ã‚¿ã‚¹ãƒˆã‚¢ãŒæä¾›ã™ã‚‹ã‚ˆã†ãªå¤šæ§˜ãªæ¤œç´¢æ©Ÿèƒ½ã¯æŒãŸãªã„ã‚ã‘ ã§ã™ï¼Žã¨ã„ã†ã‚ã‘ã§ï¼ŒGAE ã®ãƒ‡ãƒ¼ã‚¿ã‚¹ãƒˆã‚¢ã‚’å®Ÿç¾ã™ã‚‹ãŸã‚ã«ï¼ŒBigtable ãŒã©ã®ã‚ˆã†ã«ä½¿ã‚ã‚Œã¦ã„ã‚‹ã®ã‹ã‚’è€ƒãˆã¦ã¿ã¾ã—ãŸï¼Ž # ã“ã®ä»¶ã«ã¤ã„ã¦ï¼Œã‚‚ã—
hirose31 2008/04/23
è€ƒå¯Ÿ

google app engine

bigtable
ãƒªãƒ³ã‚¯
ã‚«ã‚¤ãƒãƒ„ãƒ‹ãƒƒã‚(2008-02-12)
_ [ã‚½ãƒ•ãƒˆã‚¦ã‚§ã‚¢] hadoop + hyperta ble å…ˆæ—¥è©±é¡Œã«ãªã£ã¦ã„ãŸHyperta bleãªã®ã§ã™ãŒã€ã¾ã•ã«ã“ã†ã„ã†ã‚‚ã®ãŒæ¬²ã—ã„ãªãƒ¼ã€ã¨æ€ã£ã¦flaredã¨ã‹ã‚’ä½œã‚Šå§‹ã‚ã¦ã¿ãŸã‚ã‘ãªã®ã§èˆˆå‘³æ´¥ã€…çœºã‚ã¦ã¿ã¾ã—ãŸ...ã€‚ãã®ã†ã¡4å°ãã‚‰ã„ã«å…¥ã‚Œã¦è² è·ã‹ã‘ã¦ã¿ãŸã„ã®ã§ã™ãŒã¾ã ã¡ã‚‡ã£ã¨æ™‚é–“ãŒå–ã‚Œãªã„ã®ã§ã¾ãšã¯ãƒ‰ã‚ãƒ¥ãƒ¡ãƒ³ãƒˆã‚’çœºã‚ã¦ã¿ã¾ã—ãŸã€‚ ãã‚‚ãã‚‚Hyperta bleãŒã§ãã‚‹ã“ã¨ã€ã¨ã„ã†ã®ã¯ã‚ˆãã‚ã‚‹(?)key-valueåˆ†æ•£ã‚¹ãƒˆãƒ¬ãƒ¼ã‚¸ã§ã™(ãŒã€å˜ç´”ã«keyã‹ã‚‰valueã‚’å¼•ãã¨ã„ã†ã‚ªãƒšãƒ¬ãƒ¼ã‚·ãƒ§ãƒ³ã‚ˆã‚Šã¯è¤‡é›‘ã‚ã‚‹ã„ã¯ç´°ã‹ã„å‡¦ç†ãŒå¯èƒ½ãªã‚ˆã†ã§ã™)ã€‚ ãã®Hyperta bleã¯ã€Range Serverã¨ã‚ˆã°ã‚Œã‚‹å®Ÿéš›ã®ãƒ‡ãƒ¼ã‚¿ã‚’æ ¼ç´ã™ã‚‹nå€‹ã®ã‚µãƒ¼ãƒãƒ—ãƒã‚»ã‚¹ã¨ã€Master Serverã¨ã‚ˆã°ã‚Œã‚‹1ã¤ã®ã‚³ãƒ³ãƒˆãƒãƒ¼ãƒ«ãƒ—ãƒã‚»ã‚¹ã€ãã—ã¦Hyperspaceã¨ã„ã†n + 2å€‹ã®ãƒ—ãƒã‚»ã‚¹ã«ã‚ˆã£ã¦æˆã‚Šç«‹ã¤ã®ã
hirose31 2008/02/15
bigtable

Hypertable

flared
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
hirose31 2008/02/09
bigtable
ãƒªãƒ³ã‚¯
Googleã®èª‡ã‚‹å·¨å¤§ãƒ‡ãƒ¼ã‚¿ãƒ™ãƒ¼ã‚¹BigTableã®ã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹ã‚¯ãƒãƒ¼ãƒ³ã€ŒHypertableã€
Googleã®ã‚ã®å·¨å¤§ãªæ¤œç´¢ã‚·ã‚¹ãƒ†ãƒ ãªã©ã‚’æ”¯ãˆã‚‹ãƒ‡ãƒ¼ã‚¿ãƒ™ãƒ¼ã‚¹ã€ŒBigTableã€ã¯æ›¸ãè¾¼ã¿ãŒæ¯Žç§’700MBã€èªã¿è¾¼ã¿ãŒæ¯Žç§’18GBã¨ã„ã†åŒ–ã‘ç‰©ã‚·ã‚¹ãƒ†ãƒ ãªã®ã§ã™ãŒã€ãã‚Œã®ã‚¯ãƒãƒ¼ãƒ³ã‚’ä½œã‚‹ãƒ—ãƒã‚¸ã‚§ã‚¯ãƒˆã€ŒHyperta bleã€ã¨ã„ã†ã®ãŒã‚ã‚‹ã‚ˆã†ã§ã™ã€‚æ—¢ã«ãƒãƒ¼ã‚¸ãƒ§ãƒ³ãŒä»Šå¹´ã®2æœˆ4æ—¥ã«ã€Œ0.9 Alphaã€ã¾ã§åˆ°é”ã—ã¦ãŠã‚Šã€è¶…å·¨å¤§ãªåˆ†æ•£ãƒ‡ãƒ¼ã‚¿ãƒ™ãƒ¼ã‚¹ã‚’è€ƒãˆã¦ã„ã‚‹äººã€…ã‹ã‚‰æ³¨ç›®ã•ã‚Œã¦ã„ã‚‹ã‚ˆã†ã§ã™ã€‚ è©³ç´°ã¯ä»¥ä¸‹ã‹ã‚‰ã€‚ Hyperta ble: An Open Source, High Performance, Scala ble Database http://www.hyperta ble.org/ ãƒ€ã‚¦ãƒ³ãƒãƒ¼ãƒ‰ã¯ä»¥ä¸‹ã‹ã‚‰ã€‚åŒæ¢±ã•ã‚Œã¦ã„ã‚‹ãƒ†ã‚ã‚¹ãƒˆãƒ•ã‚¡ã‚¤ãƒ«ã«ã‚¤ãƒ³ã‚¹ãƒˆãƒ¼ãƒ«æ–¹æ³•ãªã©ãŒè¨˜ã•ã‚Œã¦ã„ã¾ã™ã€‚ Download Hyperta ble å®Ÿéš›ã®ãƒ†ã‚¹ãƒˆçµæžœãªã©ã¯ä»¥ä¸‹ã«æ›¸ã„ã¦ã‚ã‚Šã¾ã™ã€‚ PerformanceTestAO
hirose31 2008/02/09
bigtable
ãƒªãƒ³ã‚¯
Bigtable - odz buffer
Google ã®å·¥è—¤ã•ã‚“*1ãŒå¤§å¦ã«æ¥ã¦è¬›æ¼”ã™ã‚‹ã¨ã„ã†ã®ã§èžã„ã¦ããŸã€‚ Bigtable ã®è©±ã§çµæ§‹é¢ç™½ã‹ã£ãŸã®ã ãŒã€ã¾ãã€ãªã‚“ã‹ã•ã™ãŒå¤§è¦æ¨¡åˆ†æ•£ã‚·ã‚¹ãƒ†ãƒ ã¨ã„ã†ã‹ã€æ›¸ãè¾¼ã¿æ€§èƒ½ 700MB/secã€èªã¿è¾¼ã¿æ€§èƒ½ 18GB/sec ã¯ç„¡ã„ã‚ã¨æ€ã£ãŸã‚ˆã€‚ *1:MeCab ã¨ã‹ CaboCha ã¨ã‹ã„ã‚ã„ã‚ä½œã£ã¦ã„ã‚‹äººã§ã‚‚ã‚ã‚‹
hirose31 2008/01/26
>æ›¸ãè¾¼ã¿æ€§èƒ½ 700MB/secã€èªã¿è¾¼ã¿æ€§èƒ½ 18GB/sec

bigtable
ãƒªãƒ³ã‚¯
Hypertable is a New BigTable Clone that Runs on HDFS orÂ KFS - High Scalability -
Update 3: Presentation from the NoSQL conference: slides, video 1, video 2. Update 2: The folks at Hyperta ble would like you to know that Hyperta ble is now officially sponsored by Baidu, Chinaâ€™s Leading Search Engine. As a sponsor of Hyperta ble, Baidu has committed an industrious team of engineers, numerous servers, and supportÂ resources to improve the quality and development of the open source te
hirose31 2008/01/17
C++ã§æ›¸ã‹ã‚ŒãŸBigTableã‚¯ãƒãƒ¼ãƒ³ã‚‰ã—ã„ï¼HDFS: http://lucene.apache.org/hadoop/ ï¼KFS: http://kosmosfs.sourceforge.net/

bigtable

Hypertable
ãƒªãƒ³ã‚¯
GoogleÂ Architecture - High Scalability -
Update 2: Sorting 1 PB with MapReduce. PB is not peanut-butter-and-jelly misspelled. It's 1 petabyte or 1000 terabytes or 1,000,000 gigabytes. It took six hours and two minutes to sort 1PB (10 trillion 100-byte records) on 4,000 computers and the results were replicated thrice on 48,000 disks. Update: Greg Linden points to a new Google article MapReduce: simplified data processing on large cluster
hirose31 2008/01/13
google

architecture

mapreduce

bigtable

googlefs
ãƒªãƒ³ã‚¯
1