Twitterã§ã¯åºæ¬çã«ãã¡ã¤ã«ã¯LZOå§ç¸®ãã¦ããããã§ï¼ 3,4åã®ã¹ãã¬ã¼ã¸ã®ç¯ç´ åå²å¯è½ CPUã¯å°ããã使ããªã IOãã¦ã³ãã®ã¸ã§ãã¯3,4åã®æ§è½åä¸ ãªã©ã®ã¡ãªãããããã¨è¨ã£ã¦ãã¾ãï¼ããã¯ä½¿ããªãæã¯ãªãã¨ãããã¨ã§è©¦ãã¦ã¿ã¾ããï¼ clouderaã®ãã®ããã°è¨äºãåèã«ãã¦é²ãã¾ãï¼ code.google.com/p/hadoop-gpl-compressionãããã¾ããï¼Twitterãå ¬éãã¦ããåå²å¯è½ãªã®ã使ãã¾ãï¼ http://github.com/kevinweil/hadoop-lzo ä»åã®ç°å¢ã¯clouderaã®amiããã¼ã¹ã«ãã¾ããï¼ cloudera-ec2-hadoop-images/cloudera-hadoop-fedora-20090623-x86_64 ami-2359bf4 CDH3ã§ï¼hadopoã®ãã¼ã¸ã§ã³ã¯
æè¿å§ç¸®ãã¡ã¤ã«ã®é度ã«ã¤ãã¦æ°ã«ãªãã®ã§ããããã調ã¹ã¦ã¿ãã¨ãå§ç¸®çã¯ä½ãããé度ã¯çéã ã¨è¨ããã¦ããLZOã¨è¨ãã®ãããã¿ããã ã HTTPã®å§ç¸®ã«ã使ããã¦ããGZIPã¯çµæ§ãªã¼ãã¼ãããå°ããã¨æã£ã¦ããã®ã ããå®éã«LZOãJavaã®JNIçµç±ã§å¼ã³åºãJavaå®è£ ãSeabassNativeIOã«è¿½å ãã¦ãããããã®é度ãéã£ã¦ã¿ãã ã¡ãªã¿ã«GZIPå§ç¸®è§£åã¯ãjava.util.zip.GZIPInputStream,java.util.zip.GZIPOutputStreamã§å¦çããã ãããããããããjava.io.ByteArrayInputStream,java.io.ByteArrayOutputSteamããã¾ãã¦å¦çããã private static final int LEN = 20 ; /** * @param args */ public s
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}}