µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
Invariant µ¥ÀÌÅ͸¦ ÀÌ¿ëÇÑ ¼øȯ󸮸¦ Áö¿øÇÏ´Â MapReduce ±â¹Ý ºÐ»ê µ¥ÀÌÅÍ Ã³¸® ½Ã½ºÅÛ |
¿µ¹®Á¦¸ñ(English Title) |
MapReduce based Distributed Data Processing System Supporting Iterative Processing using Invariant Data |
ÀúÀÚ(Author) |
À±¹Î
±èÇüÀÏ
È«½ÂÅÂ
ÃÖ¹®Ã¶
Á¶Èñ½Â
ÃÖµ¿ÈÆ
ÀåÀç¿ì
Min Yoon
Hyeong-il Kim
Seung-tae Hong
Moon-cheol Choi
Heeseung Jo
Donghoon Choi
Jae-Woo Chang
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 29 NO. 03 PP. 0013 ~ 0026 (2013. 12) |
Çѱ۳»¿ë (Korean Abstract) |
ÃÖ±Ù IT ºÐ¾ß¿¡¼ ÀÎÅͳÝÀ» ±â¹ÝÀ¸·Î IT ÀÚ¿øµéÀ» ¼ºñ½º ÇüÅ·ΠÁ¦°øÇϴ Ŭ¶ó¿ìµå ÄÄ, Ç»Æÿ¡ ´ëÇÑ °ü½ÉÀÌ Áõ´ëµÇ°í ÀÖÀ¸¸ç, ÀÌ¿¡ µû¶ó ´ë±Ô¸ð µ¥ÀÌÅ͸¦ ¼ö¸¹Àº ¼¹öµé¿¡ ºÐ»ê ÀúÀåÇÏ°í °ü¸®Çϱâ À§ÇÑ ºÐ»ê µ¥ÀÌÅÍ Ã³¸® ±â¹ý¿¡ ´ëÇÑ ¿¬±¸°¡ È°¹ßÈ÷ ÁøÇàµÇ°í ÀÖ´Ù. ÇÏÁö¸¸, Hadoop ¹× HaLoop °ú °°Àº ±âÁ¸ ¿¬±¸µéÀº ¹Ýº¹ÀûÀÎ ¼öÇàÀ» À§ÇÑ ¼øȯ ó¸® ÀÛ¾÷¿¡ ´ëÇØ È¿À²ÀûÀ¸·Î Áö¿øÇÏÁö ¸øÇÏ´Â ¹®Á¦Á¡ÀÌ Á¸ÀçÇÑ´Ù. À̸¦ À§ÇØ º» ³í¹®¿¡¼´Â ºñÀü¹® °³¹ßÀÚ¸¦ À§ÇÑ MapReduce ±â¹Ý ¼øȯ ó¸® ±â¹ýÀ» Á¦¾È(ÀÌÇÏ Hadoop X)ÇÑ´Ù. Á¦¾ÈÇÏ´Â ±â¹ýÀº ºñÀü¹® °³¹ßÀÚ°¡ È¿À²ÀûÀ¸·Î ¼øȯ ó¸® ÀÀ¿ëÀ» »ç¿ëÇÒ ¼ö ÀÖµµ·Ï, invariant µ¥ÀÌÅÍÀÇ ÀÚµ¿ °ËÃâ ±â´ÉÀ» ¼öÇàÇÑ´Ù. ¶ÇÇÑ, Ŭ·¯½ºÅÍ È¯°æÀÇ °¡¿ë ¸Þ¸ð¸®¸¦ È°¿ëÇÑ In-memory ij½ÌÀ» ¼öÇàÇÔÀ¸·Î½á, ¼øȯ ó¸® °úÁ¤¿¡¼ÀÇ Ä³½Ì ¼º´ÉÀ» Çâ»ó½ÃŲ´Ù. ¸¶Áö¸·À¸·Î ´Ù¾çÇÑ ¼øȯ ó¸® ÀÀ¿ëÀ» ÅëÇÑ ¼º´ÉÆò°¡¸¦ ¼öÇàÇÔÀ¸·Î½á, Á¦¾ÈÇÏ´Â ½Ã½ºÅÛÀÌ ±âÁ¸ ½Ã½ºÅÛº¸´Ù Çâ»óµÈ ¼º´ÉÀ» º¸ÀÓÀ» ÀÔÁõÇÑ´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
Recently, interest in cloud computing which provides IT resources as service form in IT field is increasing. As a result, much research has been done on the distributed data processing that store and manage a large amount of data in many servers. However, the existing systems, such as Hadoop and HaLoop, have problems that do not provide iterative processing efficiently. Threfore, in this paper, we propose a new iterative processing system based on MapReduce. Our proposed system supoort automatic invariant data detection scheme for the layman who does not know MapReduce system. And in-memory caching algorithm in our system can improve the performance for iterative processing. Finally, we show from performance analysis that the proposed scheme greatly improve the performance of iterative processing than the existing systems.
|
Å°¿öµå(Keyword) |
Ŭ¶ó¿ìµå ÄÄÇ»ÆÃ
´ë¿ë·® µ¥ÀÌÅÍ
ºÐ»ê ó¸®
MapReduce
¼øȯ ó¸® ½Ã½ºÅÛ
Invariant µ¥ÀÌÅÍ ÀÚµ¿ °ËÃâ
Cloud Computing
Big Data
Distributed Processing
MapReduce
Iterative Processing System
Automatic Invariant Data Detection
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|