{"id":2148,"date":"2014-04-02T09:37:17","date_gmt":"2014-04-02T01:37:17","guid":{"rendered":"http:\/\/www.banbanli.net\/blog\/?p=2148"},"modified":"2014-04-02T09:37:58","modified_gmt":"2014-04-02T01:37:58","slug":"%e8%bd%89%e8%b2%bc%e9%9b%b2%e7%ab%af%e9%81%8b%e7%ae%97%e5%b9%b3%e5%8f%b0-hadoop","status":"publish","type":"post","link":"http:\/\/www.banbanli.net\/blog\/?p=2148","title":{"rendered":"[\u8f49\u8cbc]\u96f2\u7aef\u904b\u7b97\u5e73\u53f0\u2014Hadoop"},"content":{"rendered":"<p>http:\/\/www.cc.ntu.edu.tw\/chinese\/epaper\/0011\/20091220_1106.htm<\/p>\n<p>&nbsp;<\/p>\n<p>\u4f5c\u8005\uff1a\u5468\u79c9\u8abc \/ \u81fa\u7063\u5927\u5b78\u8a08\u7b97\u6a5f\u53ca\u8cc7\u8a0a\u7db2\u8def\u4e2d\u5fc3\u4f5c\u696d\u7ba1\u7406\u7d44\u78a9\u58eb\u5f8c\u7814\u7a76\u4eba\u54e1<\/p>\n<hr size=\"1\" \/>\n<p>\u96f2\u7aef\u904b\u7b97\u662f\u8cc7\u6599\u4e2d\u5fc3\u56e0\u61c9\u7db2\u8def\u4e0a\u8cc7\u8a0a\u66b4\u589e\u800c\u63d0\u51fa\u7684\u670d\u52d9\u53ca\u7ba1\u7406\u601d\u7dad\uff0c\u8cc7\u8a0a\u670d\u52d9\u63d0\u4f9b\u8005\u6295\u5165\u8cc7\u6e90\u9032\u884c\u96f2\u7aef\u904b\u7b97\u7684\u670d\u52d9\u53ca\u67b6\u69cb\u958b\u767c\uff0cGoogle\u53ef\u8aaa\u662f\u6700\u5927\u91cf\u4f7f\u7528\u96f2\u7aef \u904b\u7b97\u7684\u7d44\u7e54\u4e4b\u4e00\u3002Hadoop\u5c31\u662f\u7531Google\u96f2\u7aef\u67b6\u69cb\u5f97\u5230\u555f\u767c\u800c\u958b\u59cb\u7684\u958b\u653e\u539f\u59cb\u78bc\u8a08\u5283\uff0c\u76ee\u524d\u6709\u8a31\u591a\u7d44\u7e54\u53c3\u8207Hadoop\u7684\u7814\u7a76\u958b\u767c\uff0c\u4e26\u4ee5Hadoop \u505a\u70ba\u96f2\u7aef\u904b\u7b97\u7684\u5e73\u53f0\u3002<\/p>\n<p><b>\u524d\u8a00<br \/>\n<\/b>\u96a8\u8457\u7db2\u969b\u7db2\u8def (Internet) \u7684\u767c\u5c55\uff0c\u53caweb2.0\u6982\u5ff5\u88ab\u63d0\u51fa\uff0c\u7db2\u8def\u4f7f\u7528\u8005\u7684\u884c\u70ba\u4e5f\u7531\u55ae\u7d14\u7684\u700f\u89bd\u8f49\u8b8a\u70ba\u5275\u4f5c\u8207\u5206\u4eab\uff1b\u53e6\u5916\uff0c\u884c\u52d5\u5f0f\u7684\u8cc7\u8a0a\u8a2d\u5099\u4e5f\u8d8a\u4f86\u8d8a\u591a\uff0c\u70ba\u4e86\u65b9\u4fbf\u5206\u4eab\u53ca\u53d6\u7528\uff0c\u4f7f\u7528\u8005\u5011\u628a\u8cc7\u6599\u5f9e\u500b\u4eba\u7684\u96fb\u8166\u4e2d\u8f49\u79fb\u5230web\u670d\u52d9\u63d0\u4f9b\u8005\u7684\u8cc7\u6599\u4e2d\u5fc3 (Data Center)\uff1b\u800c\u670d\u52d9\u63d0\u4f9b\u8005\u70ba\u4e86\u63d0\u4f9b\u66f4\u7a69\u5b9a\u66f4\u8fc5\u901f\u7684\u670d\u52d9\uff0c\u4e5f\u9700\u8981\u4e00\u500b\u65b0\u7684\u670d\u52d9\u67b6\u69cb\uff0c\u5c07\u904b\u7b97\u8cc7\u6e90\u53ca\u5132\u5b58\u7a7a\u9593\u66f4\u6709\u6548\u7387\u7684\u5229\u7528\uff0c\u540c\u6642\u63d0\u4f9b\u670d\u52d9\u958b\u767c\u4eba\u54e1\u66f4\u4fbf\u5229\u7684\u958b\u767c\u74b0\u5883\u3002<!--more--><\/p>\n<p>\u96f2\u7aef\u904b\u7b97 (Cloud Computing) \u5c31\u662f\u5c07\u524d\u8ff0\u6240\u6709\u7684\u9700\u6c42\u6574\u5408\u5728\u4e00\u8d77\u7684\u6982\u5ff5\uff0c\u4e00\u500b\u9762\u5411\u662f\u8b93\u4f7f\u7528\u8005\u4ee5\u66f4\u52a0\u4fbf\u5229\u7684\u65b9\u5f0f\u4f7f\u7528\u53ca\u53d6\u5f97\u670d\u52d9\uff0c\u751a\u81f3\u7528\u6700\u7c21\u55ae\u7684\u65b9\u5f0f\u958b\u767c\u65b0\u7684\u670d\u52d9\u3002\u96a8\u8457\u5404\u7a2e\u96f2\u7aef\u670d\u52d9\u7522\u751f\uff0c\u5c0d \u65bc\u904b\u7b97\u80fd\u529b\u53ca\u5132\u5b58\u7a7a\u9593\u7684\u9700\u6c42\uff0c\u4e5f\u6703\u9a5a\u4eba\u5730\u6210\u9577\uff0c\u56e0\u6b64\u96f2\u7aef\u904b\u7b97\u7684\u53e6\u4e00\u500b\u9762\u5411\u5c31\u662f\u6574\u5408\u7d44\u7e54\u5167\u90e8\u904b\u7b97\u8cc7\u6e90\uff0c\u4ee5\u6700\u6709\u6548\u7387\u3001\u6700\u6613\u65bc\u7ba1\u7406\u7684\u65b9\u5f0f\uff0c\u63d0\u4f9b\u96f2\u7aef\u670d\u52d9\u7a69\u5b9a\u7684\u904b \u7b97\u53ca\u5132\u5b58\u80fd\u91cf\u3002<br \/>\n\u4ee5Google\u70ba\u4f8b\uff0c\u8a31\u591a\u670d\u52d9\u90fd\u4ee5\u96f2\u7aef\u904b\u7b97\u7684\u5f62\u5f0f\u63a8\u51fa\uff0c\u8b93\u4f7f\u7528\u8005\u96a8\u6642\u53ef\u4ee5\u53d6\u5f97\u81ea\u5df1\u7684\u8cc7\u6599\uff0c\u4e5f\u80fd\u5920\u900f\u904e\u7db2\u8def\u8ddf\u5176\u4ed6\u4eba\u5206\u4eab\uff1b\u9084\u63d0\u4f9b\u4e86\u76f8\u7576\u4fbf\u5229\u7684\u958b\u767c\u74b0\u5883\uff0c\u5982 Google App. Engine\u63d0\u4f9b\u4e86\u4ecb\u9762\u548c\u514d\u8cbb\u7684\u904b\u7b97\u53ca\u5132\u5b58\u8cc7\u6e90\uff0c\u8b93\u4f7f\u7528\u8005\u958b\u767c\u5404\u7a2e\u6709\u8da3\u7684web\u670d\u52d9\u3002\u4f46\u9019\u4e9b\u670d\u52d9\u9700\u8981\u5341\u5206\u53ef\u89c0\u7684\u904b\u7b97\u80fd\u529b\u548c\u4f7f\u7528\u8005\u8cc7\u6599\u7684\u5132\u5b58\u7a7a\u9593\uff0c\u56e0 \u6b64\uff0cGoogle\u958b\u767c\u4e86\u8a31\u591a\u96f2\u7aef\u904b\u7b97\u7684\u6280\u8853\u8207\u67b6\u69cb\uff0c\u5982MapReduce\u4ee5\u5206\u6563\u5f0f\u904b\u7b97\u63d0\u4f9b\u6574\u5408\u7684\u904b\u7b97\u8cc7\u6e90\u53ca\u6e1b\u5c11\u904b\u7b97\u6642\u9593\u3001Google File System\u5c07\u5927\u91cf\u800c\u5206\u6563\u7684\u5132\u5b58\u7a7a\u9593\u6574\u5408\u70ba\u4e00\u500b\u53ef\u9760\u7684\u5132\u5b58\u5a92\u4ecb\u3001BigTable\u63d0\u4f9b\u9ad8\u6548\u7387\u7684\u5206\u6563\u5f0f\u8cc7\u6599\u5eab\u3002\u9019\u4e9b\u6280\u8853\u53ca\u67b6\u69cb\u90fd\u6709\u4e00\u500b\u7279\u9ede\uff0c\u5c31\u662f\u8b93\u670d\u52d9\u958b\u767c \u4eba\u54e1\u4e0d\u7528\u8003\u616e\u5728\u9019\u4e9b\u5206\u6563\u5f0f\u7cfb\u7d71\u4e0a\u8cc7\u6599\u8981\u600e\u9ebc\u653e\u7f6e\u3001\u904b\u7b97\u8981\u600e\u9ebc\u5207\u5272\uff0c\u53ea\u9700\u8981\u5c08\u6ce8\u5728\u670d\u52d9\u7684\u958b\u767c\u5c31\u53ef\u4ee5\u4e86\uff0c\u800c\u8cc7\u6599\u8207\u904b\u7b97\u7684\u5207\u5272\u53ca\u5206\u6563\u5c31\u4ea4\u7d66\u96f2\u7aef\u904b\u7b97\u7684\u67b6\u69cb\u4f86\u8655 \u7406\uff0c\u53ef\u8aaa\u662f\u5927\u5927\u589e\u52a0\u4e86\u958b\u767c\u670d\u52d9\u7684\u901f\u5ea6\u3002<\/p>\n<p><b>Hadoop\u8a08\u5283<br \/>\n<\/b>Hadoop\u662fApache\u8edf\u9ad4\u57fa\u91d1\u6703 (Apache Software Foundation) \u5e95\u4e0b\u7684\u958b\u653e\u539f\u59cb\u78bc\u8a08\u5283 (Open source project)\uff0c\u6700\u521d\u662f\u505a\u70baNutch\u9019\u500b\u958b\u653e\u539f\u59cb\u78bc\u7684\u641c\u5c0b\u5f15\u64ce\u7684\u4e00\u90e8\u4efd\u3002Hadoop\u662f\u4ee5java\u5beb\u6210\uff0c\u53ef\u4ee5\u63d0\u4f9b\u5927\u91cf\u8cc7\u6599\u7684\u5206\u6563\u5f0f\u904b\u7b97\u74b0\u5883\uff0c\u800c\u4e14 Hadoop\u7684\u67b6\u69cb\u662f\u7531Google\u767c\u8868\u7684BigTable\u53caGoogle File System\u7b49\u6587\u7ae0\u63d0\u51fa\u7684\u6982\u5ff5\u5be6\u505a\u800c\u6210\uff0c\u6240\u4ee5\u8ddfGoogle\u5167\u90e8\u4f7f\u7528\u7684\u96f2\u7aef\u904b\u7b97\u67b6\u69cb\u76f8\u4f3c\u3002\u76ee\u524dYahoo!\u53caCloudera\u7b49\u516c\u53f8\u90fd\u6709\u958b\u767c\u4eba\u54e1\u6295\u5165 Hadoop\u7684\u958b\u767c\u5718\u968a\uff0c\u4e5f\u6709\u5c07\u8fd1\u4e00\u767e\u500b\u516c\u53f8\u6216\u7d44\u7e54\u516c\u958b\u8868\u793a\u4f7f\u7528Hadoop\u505a\u70ba\u96f2\u7aef\u904b\u7b97\u5e73\u53f0\uff0cGoogle\u53caIBM\u4e5f\u4f7f\u7528Hadoop\u5e73\u53f0\u70ba\u6559\u80b2\u5408\u4f5c\u74b0 \u5883\u3002<br \/>\nHadoop\u4e2d\u5305\u62ec\u8a31\u591a\u5b50\u8a08\u5283\uff0c\u5176\u4e2dHadoop MapReduce\u5982\u540cGoogle MapReduce\uff0c\u63d0\u4f9b\u5206\u6563\u5f0f\u904b\u7b97\u74b0\u5883\u3001Hadoop Distributed File System\u5982\u540cGoogle File System\uff0c\u63d0\u4f9b\u5927\u91cf\u5132\u5b58\u7a7a\u9593\u3001HBase\u662f\u4e00\u500b\u985e\u4f3c BigTable \u7684\u5206\u6563\u5f0f\u8cc7\u6599\u5eab (\u898b\u8868\u4e00)\uff0c\u9084\u6709\u5176\u4ed6\u90e8\u4efd\u53ef\u7528\u4f86\u5c07\u9019\u4e09\u500b\u4e3b\u8981\u90e8\u4efd\u9023\u7d50\u5728\u4e00\u8d77\uff0c\u65b9\u4fbf\u63d0\u4f9b\u6574\u5408\u7684\u96f2\u7aef\u670d\u52d9\u3002<\/p>\n<p><img decoding=\"async\" alt=\"\" src=\"http:\/\/www.cc.ntu.edu.tw\/chinese\/epaper\/0011\/200912201106001.gif\" border=\"0\" \/><\/p>\n<p><b>MapReduce<br \/>\n<\/b>MapReduce\u662f\u4e00\u500b\u5206\u6563\u5f0f\u7a0b\u5f0f\u6846\u67b6\uff0c\u8b93\u670d\u52d9\u958b\u767c\u8005\u53ef\u4ee5\u5f88\u7c21\u55ae\u7684\u64b0\u5beb\u7a0b\u5f0f\uff0c\u5229\u7528\u5927\u91cf\u7684\u904b\u7b97\u8cc7\u6e90\uff0c\u52a0\u901f\u8655\u7406\u9f90\u5927\u7684\u8cc7\u6599\u91cf\uff0c\u4e00\u500bMapReduce\u7684\u904b\u7b97\u5de5\u4f5c\u53ef\u4ee5\u5206\u6210\u5169\u500b\u90e8\u4efd\u2014Map\u548cReduce\uff0c\u5927\u91cf\u7684\u8cc7\u6599\u5728\u904b\u7b97\u958b\u59cb\u7684\u6642\u5019\uff0c\u6703\u88ab\u7cfb\u7d71\u8f49\u63db\u6210\u4e00\u7d44\u7d44 (key, value) \u7684\u5e8f\u5c0d\u4e26\u81ea\u52d5\u5207\u5272\u6210\u8a31\u591a\u90e8\u4efd\uff0c\u5206\u5225\u50b3\u7d66\u4e0d\u540c\u7684Mapper\u4f86\u8655\u7406\uff0cMapper\u8655\u7406\u5b8c\u6210\u5f8c\u4e5f\u8981\u5c07\u904b\u7b97\u7d50\u679c\u6574\u7406\u6210\u4e00\u7d44\u7d44 (key, value) \u7684\u5e8f\u5c0d\uff0c\u518d\u50b3\u7d66Reducer\u6574\u5408\u6240\u6709Mapper\u7684\u7d50\u679c\uff0c\u6700\u5f8c\u624d\u80fd\u5c07\u6574\u9ad4\u7684\u7d50\u679c\u8f38\u51fa (\u898b\u5716\u4e00)\u3002<\/p>\n<p><img decoding=\"async\" alt=\"\" src=\"http:\/\/www.cc.ntu.edu.tw\/chinese\/epaper\/0011\/200912201106002.jpg\" border=\"0\" \/><br \/>\n\u518d\u66f4\u4ed4\u7d30\u5730\u4ecb\u7d39\u6d41\u7a0b\u4e2d\u6bcf\u4e00\u6b65\u7684\u7d30\u7bc0\uff0c\u4e00\u958b\u59cb\u9700\u8981\u5efa\u7acb\u4e00\u500bJobConf\u985e\u5225\u7684\u7269\u4ef6\uff0c\u7528\u4f86\u8a2d\u5b9a\u904b\u7b97\u5de5\u4f5c\u7684\u5167\u5bb9\uff0c\u5982 setMapperClass\/setReducerClass\u8a2d\u5b9a Mapper\u53caReducer \u7684\u985e\u5225\uff0csetInputFormat\/setOutputFormat \u8a2d\u5b9a\u8f38\u51fa\u8f38\u5165\u8cc7\u6599\u7684\u683c\u5f0f\uff0c\u00a0 setOutputKeyClass \/ setOutputValueClass \u8a2d\u5b9a\u8f38\u51fa\u8cc7\u6599\u7684\u985e\u578b\uff0c\u8a2d\u5b9a\u5b8c\u6210\u5f8c\uff0c\u4f9d\u8a2d\u5b9a\u5167\u5bb9\u63d0\u4ea4\u904b\u7b97\u5de5\u4f5c\u3002\u8cc7\u6599\u4f86\u6e90\u6703\u4f9dInputFormat\u7684\u8a2d\u5b9a\u53d6\u5f97\uff0c\u4e26\u5206\u5272\u8f49\u63db\u70ba\u4e00\u7d44\u7d44\u7684 (key, value) \u5e8f\u5c0d\uff0c\u4ea4\u7531\u4e0d\u540c\u7684Mapper\u540c\u6642\u9032\u884c\u904b\u7b97\uff0cMapper\u8981\u5c07\u904b\u7b97\u7684\u7d50\u679c\u8f38\u51fa\u70ba\u4e00\u7d44\u7d44(key, value) \u5e8f\u5c0d\uff0c\u4e5f\u7a31\u70ba\u4e2d\u4ecb\u8cc7\u6599 (intermediate)\uff0c\u7cfb\u7d71\u6703\u5c07\u9019\u4e9b\u66ab\u6642\u7684\u7d50\u679c\u6392\u5e8f (sort) \u4e26\u66ab\u5b58\u8d77\u4f86\uff0c\u7b49\u5230\u6240\u6709Mapper\u7684\u904b\u7b97\u5de5\u4f5c\u7d50\u675f\u4e4b\u5f8c\uff0c\u4f9d\u7167\u4e0d\u540c\u7684key\u503c\u50b3\u9001\u7d66\u4e0d\u540c\u7684Reducer\u5f59\u6574\uff0c\u6240\u6709\u540c\u4e00key\u503c\u7684\u4e2d\u4ecb\u8cc7\u6599\u7684value\u503c\uff0c \u6703\u653e\u5728\u4e00\u500b\u5bb9\u5668 (container) \u88e1\u50b3\u7d66\u540c\u4e00\u500bReducer\u8655\u7406\uff0c\u6240\u4ee5\u5728Reducer\u4e2d\u53ef\u4ee5\u5229\u7528values.next()\u4f9d\u5e8f\u53d6\u5f97\u4e0d\u540cvalue\u503c\uff0c\u5feb\u901f\u5730\u5b8c\u6210\u7d50\u679c\u6574\u7406\uff0c\u518d\u4f9d OutputFormat\u7684\u8a2d\u5b9a\u8f38\u51fa\u70ba\u6a94\u6848\u3002<br \/>\n\u9032\u884c\u904b\u7b97\u7684Mapper\u548cReducer\u6703\u7531\u7cfb\u7d71\u6703\u81ea\u52d5\u6307\u6d3e\u4e0d\u540c\u7684\u904b\u7b97\u7bc0\u9ede\u64d4\u4efb\uff0c\u6240\u4ee5\u7a0b\u5f0f\u8a2d\u8a08\u6642\u5b8c\u5168\u4e0d\u7528\u505a\u8cc7\u6599\u548c\u904b\u7b97\u7684\u5207\u5272 (decomposition)\uff0c\u904b\u7b97\u8cc7\u6e90\u6703\u7531JobTracker\u5206\u914d\u5230\u5404\u500b\u904b\u7b97\u7bc0\u9ede\u4e0a\u7684TaskTracker\uff0c\u4e26\u6307\u6d3e\u4e0d\u540c\u7684\u7bc0\u9ede\u64d4\u4efbMapper\u548cReducer\u3002<\/p>\n<p><b>HDFS<br \/>\n<\/b>Hadoop Distributed File System (HDFS) \u5c07\u5206\u6563\u7684\u5132\u5b58\u8cc7\u6e90\u6574\u5408\u6210\u4e00\u500b\u5177\u5bb9\u932f\u80fd\u529b\u3001\u9ad8\u6548\u7387\u4e14\u8d85\u5927\u5bb9\u91cf\u7684\u5132\u5b58\u74b0\u5883\uff0c\u5728Hadoop\u7cfb\u7d71\u4e2d\u5927\u91cf\u7684\u8cc7\u6599\u548c\u904b\u7b97\u6642\u7522\u751f\u7684\u66ab\u5b58\u6a94\u6848\uff0c\u90fd\u662f\u5b58\u653e\u5728\u9019\u500b\u5206\u6563\u5f0f\u7684\u6a94\u6848\u7cfb\u7d71\u4e0a\u3002<br \/>\nHDFS\u662fmaster\/slave\u67b6\u69cb\uff0c\u7531\u5169\u7a2e\u89d2\u8272\u7d44\u6210\uff0cName node\u53cadata nodes\uff0cName node\u8ca0\u8cac\u6a94\u6848\u7cfb\u7d71\u4e2d\u5404\u500b\u6a94\u6848\u5c6c\u6027\u6b0a\u9650\u7b49\u8cc7\u8a0a (metadata, namespace) \u7684\u7ba1\u7406\u53ca\u5132\u5b58\uff1b\u800cdata node\u901a\u5e38\u7531\u6578\u4ee5\u767e\u8a08\u7684\u7bc0\u9ede\u64d4\u4efb\uff0c\u4e00\u500b\u8cc7\u6599\u6a94\u6703\u88ab\u5207\u5272\u6210\u6578\u500b\u8f03\u5c0f\u7684\u5340\u584a (block) \u5132\u5b58\u5728\u4e0d\u540c\u7684data node\u4e0a\uff0c\u6bcf\u4e00\u500b\u5340\u584a\u9084\u6703\u6709\u6578\u4efd\u526f\u672c (replica) \u5b58\u653e\u5728\u4e0d\u540c\u7bc0\u9ede\uff0c\u9019\u6a23\u7576\u5176\u4e2d\u4e00\u500b\u7bc0\u9ede\u640d\u58de\u6642\uff0c\u6a94\u6848\u7cfb\u7d71\u4e2d\u7684\u8cc7\u6599\u9084\u80fd\u4fdd\u5b58\u7121\u7f3a\uff0c\u56e0\u6b64name node\u9084\u9700\u8981\u7d00\u9304\u6bcf\u4e00\u4efd\u6a94\u6848\u5b58\u653e\u7684\u4f4d\u7f6e\uff0c\u7576\u6709\u5b58\u53d6\u6a94\u6848\u7684\u9700\u6c42\u6642\uff0c\u5354\u8abfdata node\u8ca0\u8cac\u56de\u61c9\uff1b\u800c\u6709\u7bc0\u9ede\u640d\u58de\u6642\uff0cname node\u4e5f\u6703\u81ea\u52d5\u9032\u884c\u8cc7\u6599\u7684\u642c\u9077\u548c\u8907\u88fd\u3002<br \/>\nHDFS\u96d6\u7136\u6c92\u6709\u6574\u5408\u9032Linux kernel\uff0c\u53ea\u80fd\u900f\u904eHadoop\u7684dfs shell\u9032\u884c\u6a94\u6848\u64cd\u4f5c\uff0c\u6216\u4f7f\u7528FUSE\u6210\u70baUser space\u4e0b\u7684\u6a94\u6848\u7cfb\u7d71\uff0c\u4f46Hadoop\u4e0b\u7684\u7cfb\u7d71\u90fd\u8207HDFS\u6574\u5408\uff0c\u505a\u70ba\u8cc7\u6599\u5132\u5b58\u5099\u4efd\u53ca\u5206\u4eab\u7684\u5a92\u4ecb\u3002\u5982\u524d\u9762\u63d0\u5230\u7684MapReduce\u5728\u7cfb\u7d71\u5206\u914d\u904b\u7b97\u5de5\u4f5c\u6642\uff0c\u6703\u5c07\u904b\u7b97\u5de5\u4f5c\u5206\u914d\u5230\u5b58\u653e\u6709\u904b\u7b97\u8cc7\u6599\u7684\u7bc0\u9ede\u4e0a\u9032\u884c\uff0c\u6e1b\u5c11\u5927\u91cf\u8cc7\u6599\u900f\u904e\u7db2\u8def\u50b3\u8f38\u7684\u6642\u9593\u3002<\/p>\n<p><b>HBase<br \/>\n<\/b>HBase\u662f\u67b6\u69cb\u5728HDFS\u4e0a\u7684\u5206\u6563\u5f0f\u8cc7\u6599\u5eab\uff0c\u8207\u4e00\u822c\u95dc\u806f\u5f0f\u8cc7\u6599\u5eab (relational database) \u4e0d\u540c\u3002HBase\u4f7f\u7528\u5217 (row) \u548c\u884c (column) \u70ba\u7d22\u5f15\u5b58\u53d6\u8cc7\u6599\u503c\uff0c\u56e0\u6b64\u67e5\u8a62\u7684\u6642\u5019\u6bd4\u8f03\u50cf\u5728\u4f7f\u7528map\u5bb9\u5668 (container)\uff1bHBase\u7684\u53e6\u4e00\u500b\u7279\u9ede\u662f\u6bcf\u4e00\u7b46\u8cc7\u6599\u90fd\u6709\u4e00\u500b\u6642\u9593\u6233\u8a18 (timestamp)\uff0c\u56e0\u6b64\u540c\u4e00\u500b\u6b04\u4f4d\u53ef\u4f9d\u4e0d\u540c\u6642\u9593\u5b58\u5728\u591a\u7b46\u8cc7\u6599\u3002<br \/>\n\u4e00\u500bHBase\u7684\u8cc7\u6599\u8868 (table) \u662f\u7531\u8a31\u591arow\u53ca\u6578\u500bcolumn family\u7d44\u6210\uff0c\u6bcf\u500b\u5217\u90fd\u6709\u4e00\u500brow key\u505a\u70ba\u7d22\u5f15\uff1b\u4e00\u500bcolumn family\u5c31\u662f\u4e00\u500bcolumn label\u7684\u96c6\u5408 (set)\uff0c\u88e1\u9762\u53ef\u6709\u5f88\u591a\u7d44label\uff0c\u9019\u4e9blabel\u53ef\u4ee5\u8996\u9700\u8981\u96a8\u6642\u65b0\u589e\uff0c\u800c\u4e0d\u7528\u91cd\u65b0\u8a2d\u5b9a\u6574\u500b\u8cc7\u6599\u8868 (\u898b\u8868\u4e8c)\u3002\u5728\u5b58\u53d6\u8cc7\u6599\u8868\u7684\u6642\u5019\uff0c\u901a\u5e38\u5c31\u4f7f\u7528 (\u2018row key\u2019, \u2018family:label\u2019) \u6216 (\u2018row key\u2019, \u2018family:label\u2019, \u2018timestamp\u2019) \u7684\u7d44\u5408\u53d6\u51fa\u9700\u8981\u7684\u6b04\u4f4d\u3002<\/p>\n<p><img decoding=\"async\" alt=\"\" src=\"http:\/\/www.cc.ntu.edu.tw\/chinese\/epaper\/0011\/200912201106003.gif\" border=\"0\" \/><br \/>\nHBase\u70ba\u4e86\u65b9\u4fbf\u5206\u6563\u8cc7\u6599\u548c\u904b\u7b97\u5de5\u4f5c\uff0c\u53c8\u5c07\u6574\u500b\u8cc7\u6599\u8868\u5206\u70ba\u8a31\u591aregion\uff0c\u4e00\u500bregion\u662f\u7531\u4e00\u5230\u6578\u500b\u5217\u6240\u7d44\u6210\u7684\uff0c\u53ef\u4ee5\u5206\u5225\u5b58\u653e\u5728\u4e0d\u540cHBase\u4e3b\u6a5f\u4e0a\uff0c\u9019\u4e9b\u5b58\u653eregion\u7684\u4e3b\u6a5f\u5c31\u662fregion server\uff0c\u53e6\u5916\u9084\u6709master server\u7528\u4f86\u7d00\u9304\u6bcf\u4e00\u500bregion\u5c0d\u61c9\u7684region server\uff1bmaster server\u4e5f\u6703\u81ea\u52d5\u5c07\u4e0d\u80fd\u63d0\u4f9b\u670d\u52d9\u7684region server\u4e0a\u7684region\u91cd\u65b0\u5206\u914d\u5230\u5176\u4ed6\u7684region server\u4e0a\u3002<br \/>\nHBase\u4e5f\u53ef\u4f9bMapReduce\u7684\u7a0b\u5f0f\u7576\u4f5c\u8cc7\u6599\u4f86\u6e90\u6216\u5132\u5b58\u5a92\u4ecb\uff0c\u5728HBase 0.20\u7248\u4e4b\u5f8c\u63d0\u4f9b\u4e86TableMapper\u53caTableReducer\u7684\u985e\u5225\u8b93\u7a0b\u5f0f\u4e2d\u7684Mapper\u53caReducer\u985e\u5225\u7e7c\u627f\uff0c\u53ef\u4ee5\u628aMapReuce\u4e2d\u7684 (key, value) \u66f4\u65b9\u4fbf\u5730\u5f9eHBase\u4e2d\u53d6\u51fa\u548c\u5b58\u5165\u3002<\/p>\n<p><b>Web Interface<br \/>\n<\/b>MapReduce\u7684JobTracker\u3001HDFS\u3001\u53caHBase\u90fd\u6709\u5404\u81ea\u7684web\u76e3\u63a7\u4ecb\u9762\uff0c\u53ef\u4ee5\u53ca\u6642\u89c0\u5bdf\u76ee\u524d\u6bcf\u500b\u904b\u7b97\u5de5\u4f5c\u7684\u904b\u4f5c\u60c5\u6cc1\u3001\u6a94\u6848\u7cfb\u7d71\u7684\u5bb9\u91cf\u3001\u53ca\u8cc7\u6599\u8868\u548cregion\u7684\u4f7f\u7528\u60c5\u6cc1, \u8b93\u7cfb\u7d71\u7ba1\u7406\u8005\u8f15\u9b06\u5730\u76e3\u63a7\u5927\u91cf\u8cc7\u6e90 (\u898b<b>\u5716\u4e8c\u3001\u5716\u4e09\u3001\u5716\u56db<\/b>)\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" alt=\"\" src=\"http:\/\/www.cc.ntu.edu.tw\/chinese\/epaper\/0011\/200912201106004.jpg\" width=\"544\" height=\"454\" border=\"0\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" alt=\"\" src=\"http:\/\/www.cc.ntu.edu.tw\/chinese\/epaper\/0011\/200912201106005.jpg\" width=\"544\" height=\"405\" border=\"0\" \/><br \/>\n<img decoding=\"async\" alt=\"\" src=\"http:\/\/www.cc.ntu.edu.tw\/chinese\/epaper\/0011\/200912201106006.jpg\" border=\"0\" \/><br \/>\n<b>\u7d50\u8ad6<br \/>\n<\/b>Hadoop\u662f\u76ee\u524d\u6700\u5e38\u898b\u4e14\u5be6\u969b\u904b\u7528\u5728\u5927\u898f\u6a21\u5546\u696d\u74b0\u5883\u4e0a\u7684\u96f2\u7aef\u904b\u7b97\u5e73\u53f0\u4e4b\u4e00\uff0c\u5f37\u5927\u800c\u5b8c\u6574\u7684\u57fa\u790e\u67b6\u69cb\u53ef\u4ee5\u6e1b\u5c11\u5927\u91cf\u7684\u96f2\u7aef\u67b6\u69cb\u958b\u767c\u7684\u6642\u9593\uff0c\u5927\u91cf \u90e8\u7f72\u6642\u4e5f\u76f8\u7576\u8fc5\u901f\uff0c\u4e0d\u4f46\u6709\u8a31\u591a\u91cd\u91cf\u7d1a\u7684\u96f2\u7aef\u904b\u7b97\u670d\u52d9\u63d0\u4f9b\u8005\u6b63\u5728\u4f7f\u7528\u53ca\u6295\u5165\u958b\u767c\uff0c\u4e5f\u8207Google\u7684\u96f2\u7aef\u74b0\u5883\u76f8\u4f3c\uff0c\u4f7fHadoop\u6210\u70ba\u6559\u80b2\u8a13\u7df4\u3001\u5b78\u8853\u7814\u7a76\u53ca \u96f2\u7aef\u670d\u52d9\u958b\u767c\u7684\u6700\u4f73\u5e73\u53f0\u3002<br \/>\n\u96d6\u7136\u6709Hadoop\u9019\u9ebc\u4fbf\u5229\u7684\u96f2\u7aef\u904b\u7b97\u74b0\u5883\uff0c\u53c8\u6709\u6210\u529f\u7684\u96f2\u7aef\u670d\u52d9\u53ef\u4ee5\u53c3\u8003\uff0c\u7136\u800c\u5728\u7d44\u7e54\u5167\u90e8\u5c0e\u5165\u96f2\u7aef\u904b\u7b97\u7684\u67b6\u69cb\u53ca\u6587\u5316\u6642\uff0c\u4ecd\u9700\u505a\u597d\u5145\u5206\u7684\u898f\u5283\u53ca\u6642\u7a0b\u8868\uff0c\u4e0d\u7136 \u5c07\u6703\u5f71\u97ff\u539f\u6709\u670d\u52d9\u7684\u7a69\u5b9a\u53ca\u54c1\u8cea\uff0c\u4e0d\u4f46\u4e0d\u80fd\u4eab\u53d7\u96f2\u7aef\u904b\u7b97\u5e36\u4f86\u7684\u4fbf\u5229\uff0c\u53cd\u800c\u5f92\u7136\u589e\u52a0\u7ba1\u7406\u53ca\u71df\u904b\u6210\u672c\uff0c\u4f7f\u96f2\u7aef\u904b\u7b97\u6dea\u70ba\u5931\u6557\u7684\u884c\u92b7\u540d\u8a5e\u3002<\/p>\n<p>&nbsp;<\/p>\n<p>&nbsp;<\/p>\n<p>&nbsp;<\/p>\n","protected":false},"excerpt":{"rendered":"<p>http:\/\/www.cc.ntu.edu.tw\/chinese\/epaper\/0011\/20091220_1106.htm &nbsp; \u4f5c\u8005\uff1a\u5468\u79c9\u8abc \/ \u81fa\u7063\u5927\u5b78\u8a08\u7b97\u6a5f\u53ca\u8cc7\u8a0a\u7db2\u8def\u4e2d\u5fc3\u4f5c\u696d\u7ba1\u7406\u7d44\u78a9\u58eb\u5f8c\u7814\u7a76\u4eba\u54e1 \u96f2\u7aef\u904b\u7b97\u662f\u8cc7\u6599\u4e2d\u5fc3\u56e0\u61c9\u7db2\u8def\u4e0a\u8cc7\u8a0a\u66b4\u589e\u800c\u63d0\u51fa\u7684\u670d\u52d9\u53ca\u7ba1\u7406\u601d\u7dad\uff0c\u8cc7\u8a0a\u670d\u52d9\u63d0\u4f9b\u8005\u6295\u5165\u8cc7\u6e90\u9032\u884c\u96f2\u7aef\u904b\u7b97\u7684\u670d\u52d9\u53ca\u67b6\u69cb\u958b\u767c\uff0cGoogle\u53ef\u8aaa\u662f\u6700\u5927\u91cf\u4f7f\u7528\u96f2\u7aef \u904b\u7b97\u7684\u7d44\u7e54\u4e4b\u4e00\u3002Hadoop\u5c31\u662f\u7531Google\u96f2\u7aef\u67b6\u69cb\u5f97\u5230\u555f\u767c\u800c\u958b\u59cb\u7684\u958b\u653e\u539f\u59cb\u78bc\u8a08\u5283\uff0c\u76ee\u524d\u6709\u8a31\u591a\u7d44\u7e54\u53c3\u8207Hadoop\u7684\u7814\u7a76\u958b\u767c\uff0c\u4e26\u4ee5Hadoop \u505a\u70ba\u96f2\u7aef\u904b\u7b97\u7684\u5e73\u53f0\u3002 \u524d\u8a00 \u96a8\u8457\u7db2\u969b\u7db2\u8def (Internet) \u7684\u767c\u5c55\uff0c\u53caweb2.0\u6982\u5ff5\u88ab\u63d0\u51fa\uff0c\u7db2\u8def\u4f7f\u7528\u8005\u7684\u884c\u70ba\u4e5f\u7531\u55ae\u7d14\u7684\u700f\u89bd\u8f49\u8b8a\u70ba\u5275\u4f5c\u8207\u5206\u4eab\uff1b\u53e6\u5916\uff0c\u884c\u52d5\u5f0f\u7684\u8cc7\u8a0a\u8a2d\u5099\u4e5f\u8d8a\u4f86\u8d8a\u591a\uff0c\u70ba\u4e86\u65b9\u4fbf\u5206\u4eab\u53ca\u53d6\u7528\uff0c\u4f7f\u7528\u8005\u5011\u628a\u8cc7\u6599\u5f9e\u500b\u4eba\u7684\u96fb\u8166\u4e2d\u8f49\u79fb\u5230web\u670d\u52d9\u63d0\u4f9b\u8005\u7684\u8cc7\u6599\u4e2d\u5fc3 (Data Center)\uff1b\u800c\u670d\u52d9\u63d0\u4f9b\u8005\u70ba\u4e86\u63d0\u4f9b\u66f4\u7a69\u5b9a\u66f4\u8fc5\u901f\u7684\u670d\u52d9\uff0c\u4e5f\u9700\u8981\u4e00\u500b\u65b0\u7684\u670d\u52d9\u67b6\u69cb\uff0c\u5c07\u904b\u7b97\u8cc7\u6e90\u53ca\u5132\u5b58\u7a7a\u9593\u66f4\u6709\u6548\u7387\u7684\u5229\u7528\uff0c\u540c\u6642\u63d0\u4f9b\u670d\u52d9\u958b\u767c\u4eba\u54e1\u66f4\u4fbf\u5229\u7684\u958b\u767c\u74b0\u5883\u3002<\/p>\n","protected":false},"author":3,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[51],"tags":[],"class_list":["post-2148","post","type-post","status-publish","format-standard","hentry","category-big-data"],"_links":{"self":[{"href":"http:\/\/www.banbanli.net\/blog\/index.php?rest_route=\/wp\/v2\/posts\/2148","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.banbanli.net\/blog\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.banbanli.net\/blog\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.banbanli.net\/blog\/index.php?rest_route=\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"http:\/\/www.banbanli.net\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=2148"}],"version-history":[{"count":1,"href":"http:\/\/www.banbanli.net\/blog\/index.php?rest_route=\/wp\/v2\/posts\/2148\/revisions"}],"predecessor-version":[{"id":2149,"href":"http:\/\/www.banbanli.net\/blog\/index.php?rest_route=\/wp\/v2\/posts\/2148\/revisions\/2149"}],"wp:attachment":[{"href":"http:\/\/www.banbanli.net\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=2148"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.banbanli.net\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=2148"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.banbanli.net\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=2148"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}