• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ÇÐȸÁö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ÇÐȸÁö > µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

Current Result Document : 4 / 12 ÀÌÀü°Ç ÀÌÀü°Ç   ´ÙÀ½°Ç ´ÙÀ½°Ç

ÇѱÛÁ¦¸ñ(Korean Title) ¸Þ¸ð¸® ±â¹Ý Å°-°ª ÀúÀå¼Ò¿¡¼­ÀÇ µ¥ÀÌÅÍ ¾ÐÃà ÀúÀå°ú º´·Ä ½º³À¼¦ »ý¼º ±â¹ý
¿µ¹®Á¦¸ñ(English Title) Data Compression Storage and Parallel Snapshot Generation Technique in In-memory Key-Value Stores
ÀúÀÚ(Author) ¼ºÇѽ   ¹Ú»óÇö   Hanseung Sung   Sanghyun Park  
¿ø¹®¼ö·Ïó(Citation) VOL 35 NO. 02 PP. 0033 ~ 0053 (2019. 08)
Çѱ۳»¿ë
(Korean Abstract)
½Ç½Ã°£ µ¥ÀÌÅÍ Ã³¸®¿Í ÀúÀåÀÌ ¿ä±¸µÇ´Â ȯ°æ¿¡¼­ ¸Þ¸ð¸® ±â¹Ý Å°-°ª ÀúÀå¼ÒÀÎ Redis°¡ ³Î¸® »ç¿ëµÇ°í ÀÖ´Ù. ±×·¯³ª µð½ºÅ©º¸´Ù Á¦ÇÑµÈ ÀúÀå ¿ë·®À» °¡Áø ¸Þ¸ð¸®°¡ ÁÖ ÀúÀå ÀåÄ¡·Î »ç¿ëµÇ±â ¶§¹®¿¡, ÀúÀå °¡´ÉÇÑ µ¥ÀÌÅÍÀÇ ¾çÀÌ ´Ù¼Ò Á¦ÇÑÀûÀÌ´Ù. ¶ÇÇÑ, ½Ã½ºÅÛ Àå¾Ö°¡ ¹ß»ýÇÒ °æ¿ì DRAM ÀÇ Èֹ߼ºÀ¸·Î ÀÎÇØ ÀúÀåµÈ ¸ðµç µ¥ÀÌÅÍ°¡ À¯½ÇµÇ´Â ¹®Á¦°¡ ÀÖ´Ù. À̸¦ ¹æÁöÇϱâ À§ÇÑ ¿µ¼Ó¼º ¹æ¹ýÀ» Á¦°øÇÏÁö¸¸, ¿µ¼Ó¼ºÀ» À§ÇÑ ÀÛ¾÷À¸·Î ÀÎÇØ µ¥ÀÌÅÍ Ã³¸® ¼º´ÉÀÌ ÀúÇÏµÇ°í ¸Þ¸ð¸® »ç¿ë·®ÀÌ ±ÞÁõÇÏ´Â ¹®Á¦°¡ ÀÖ´Ù. º» ³í¹®¿¡¼­´Â LZF ¾Ë°í¸®ÁòÀ» ÅëÇص¥ÀÌÅ͸¦ ¾ÐÃà ÀúÀåÇÏ¿© ¸Þ¸ð¸®ÀÇ ¿ë·® Á¦ÇÑÀ» ±Øº¹ÇÏ´Â µ¥ÀÌÅÍ ¾ÐÃà ÀúÀå ±â¹ý°ú ½º³À¼¦À» »ý¼º ½Ã ¹ß»ýÇÏ´Â ·Î±ë ºÎÇϸ¦ ¿ÏÈ­Çϱâ À§ÇØ µ¥ÀÌÅÍ º´·Ä¼ºÀ» È°¿ëÇÏ´Â º´·Ä ½º³À¼¦ »ý¼º ±â¹ýÀ» Á¦¾ÈÇÑ´Ù. º» ³í¹®¿¡¼­ Á¦¾ÈÇÏ´Â ¹æ¹ýÀ» Àû¿ëÇÑ Redis ´Â ±âÁ¸ Redis º¸´Ù ÀûÀº ¸Þ¸ð¸® »ç¿ë·®À» º¸¿´À¸¸ç, ½º³À¼¦ »ý¼º ¹× µ¥ÀÌÅÍ º¹±¸¿¡ ¼Ò¿äµÇ´Â ½Ã°£ÀÌ Å©°Ô ´ÜÃàµÇ¾úÀ½À» È®ÀÎÇÏ¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
Redis, a memory-based key-value store, is widely used in environments where real-time data processing and storage is required. However, because a memory with a limited storage capacity compared to disk is used as the primary storage device, the amount of data that can be stored is rather limited. In addition, in the event of system failure, there is a problem that all stored data are lost due to the volatility of the DRAM. Although it provides a persistence method to prevent this, there is a problem that data processing performance is degraded and memory usage is rapidly increased due to work for durability. In this paper, we propose a data compression storage technique to overcome the capacity limitation of memory by compressing and storing data through LZF algorithm and a parallel snapshot creation method that utilizes parallelism to alleviate the logging overhead that occurs when creating a snapshot. Redis using the proposed methods showed less memory usage than the existing Redis, and identified that the time required for snapshot creation and data recovery is also significantly reduced.
Å°¿öµå(Keyword) ÀÎ-¸Þ¸ð¸® Å°-°ª µ¥ÀÌÅͺ£À̽º   ·¹µð½º   µ¥ÀÌÅÍ ¾ÐÃà   ½º³À¼¦   µ¥ÀÌÅÍ º´·Ä¼º   In-memory Key-Value Database   Redis   Data Compression   Snapshot   Data Parallelis  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå