Á¤º¸°úÇÐȸ ³í¹®Áö C : ÄÄÇ»ÆÃÀÇ ½ÇÁ¦
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
¸Ê¸®µà½º±â¹Ý ½ºÆ®¸² µ¥ÀÌÅÍ Ã³¸®¸¦ À§ÇÑ ½ºÆ®¸² ÇÒ´ç ¹× ºÐÇÒ ±â¹ý |
¿µ¹®Á¦¸ñ(English Title) |
MapReduce-based Stream Assigning and Splitting Technique for Stream Data Processing |
ÀúÀÚ(Author) |
¹Ú¼öÇö
·ù¿ì¼®
È«ºÀÈñ
±ÇÁØÈ£
SooHyun Park
WooSeok Ryu
BongHee Hong
JoonHo Kwon
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 19 NO. 08 PP. 0439 ~ 0443 (2013. 08) |
Çѱ۳»¿ë (Korean Abstract) |
ÃÖ±Ù ¼¾¼ ±â±â »ç¿ëÀÇ Áõ°¡·Î ÀÎÇØ ¼¾¼·ÎºÎÅÍ ¹ß»ýÇÏ´Â µ¥ÀÌÅÍ·® ¿ª½Ã ºü¸£°Ô Áõ°¡ÇÏ°í ÀÖ´Ù. µû¶ó¼ ºòµ¥ÀÌÅÍ Ã³¸® Ç÷§ÆûÀÎ ÇϵÓÀ» È°¿ëÇÏ¿© Áõ°¡ÇÏ°í ÀÖ´Â ½ºÆ®¸² µ¥ÀÌÅ͸¦ ó¸®ÇÏ°íÀÚ ÇÑ´Ù. ±×·¯³ª ÇϵÓÀº ÀÏ°ý ¹èġó¸® Ç÷§ÆûÀ¸·Î ½ºÆ®¸² µ¥ÀÌÅÍ¿¡ ´ëÇÑ ½Ç½Ã°£ 󸮸¦ Áö¿øÇÏÁö ¾Ê´Â´Ù. µû¶ó¼ ÀÌ ³í¹®¿¡¼´Â ¿¬¼ÓÀûÀ¸·Î µé¾î¿À´Â ½ºÆ®¸² µ¥ÀÌÅ͸¦ ó¸®Çϱâ À§ÇØ ¸Ê¸®µà½º ¿Â¶óÀÎÀÇ ÆÄÀÌÇÁ¶óÀÎ ±â¹ýÀ» »ç¿ëÇÏ°í ½ºÆ®¸² ÀÔ·ÂÀ» ¹Þ±â À§ÇØ ¸Ê¸®µà½º¸¦ È®ÀåÇÏ¿´´Ù. ¶ÇÇÑ ½ºÆ®¸² µ¥ÀÌÅ͸¦ ó¸®ÇÏ¸é¼ ¹ß»ýÇÏ´Â ºÎÇÏ ¹ß»ý ¹®Á¦¸¦ ÇØ°áÇϱâ À§ÇØ ¿©·¯ ½ºÆ®¸² µ¥ÀÌÅ͸¦ ¾î¶»°Ô ´Ù¼öÀÇ Å¬·¯½ºÅÍ ³ëµå¿¡ ÇÒ´çÇÏ°í µ¥ÀÌÅ͸¦ ºÐÇÒÇÒ °ÍÀÎÁö¿¡ ´ëÇÏ¿© ¿¬±¸ÇÏ¿´´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
The increasing uses of sensors are increasing the amount of stream data. In this paper, we studied the techniques of distributed parallel processing of stream data using Hadoop. Hadoop is designed to be a batch processing system; it is not made for processing stream data. If original Hadoop is used for processing stream data, when it encountered multiple streams there will be load at a specific Hadoop Map Task. This is happening because the volume and incoming rates of stream data is vary. Therefore, we extend Hadoop's MapReduce and use MapReduce Online's pipeline for processing stream data and cope with load balancing.
|
Å°¿öµå(Keyword) |
ÇϵÓ
¸Ê¸®µà½º
¸Ê¸®µà½º ¿Â¶óÀÎ
½ºÆ®¸²Ã³¸®
ºÎÇϺлê
hadoop
MapReduce
MapReduce online
stream-proces
load-balancing
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|