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

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ³í¹®Áö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ³í¹®Áö > Çѱ¹Á¤º¸°úÇÐȸ ³í¹®Áö > Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)

Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) Å°-¹ë·ù ½ºÅä¾î¿¡¼­ÀÇ °í¼º´É I/O 󸮸¦ À§ÇÑ ¸®´ª½º I/O ÀÎÅÍÆäÀ̽º ºÐ¼®
¿µ¹®Á¦¸ñ(English Title) Analyses of Linux I/O Interfaces for High-Performance I/O Operations in Key-Value Stores
ÀúÀÚ(Author) ¼Û¿ëÁÖ   ¾ö¿µÀÍ   Yongju Song   Young Ik Eom  
¿ø¹®¼ö·Ïó(Citation) VOL 48 NO. 12 PP. 1274 ~ 1280 (2021. 12)
Çѱ۳»¿ë
(Korean Abstract)
ÃÖ±Ù °í¼º´É ½ºÅ丮Áö¸¦ È°¿ëÇϱâ À§ÇÑ ´Ù¾çÇÑ I/O ÀÎÅÍÆäÀ̽º°¡ ¿¬±¸µÇ¾î ¿ÔÀ¸¸ç, ƯÈ÷, ºñµ¿±â I/O ÀÎÅÍÆäÀ̽º°¡ ÁÖ¸ñÀ» ¹Þ°í ÀÖ´Ù. ÇÏÁö¸¸ ÀÌ´Â µ¿±â I/O ÀÎÅÍÆäÀ̽ºº¸´Ù »ó´ëÀûÀ¸·Î ÀÀ´ä½Ã°£ ³ô°í CPU ÀÚ¿øÀ» ´õ ¸¹ÀÌ »ç¿ëÇÑ´Ù´Â ´ÜÁ¡ÀÌ ÀÖ´Ù. º» ³í¹®¿¡¼­´Â °¢ I/O ÀÎÅÍÆäÀ̽ºÀÇ Æ¯¼ºÀ» ºÐ¼®ÇÏ°í, À̸¦ Å°-¹ë·ù ½ºÅä¾î¿¡ Àû¿ëÇÏ¿© ¾îÇø®ÄÉÀÌ¼Ç °èÃþ¿¡¼­ÀÇ ¼º´É ¾ç»ó¿¡ ´ëÇØ »ìÆ캻´Ù. ½ÇÇè °á°ú, ºñµ¿±âÀû I/O ÀÎÅÍÆäÀ̽º´Â µ¿±âÀû I/O ÀÎÅÍÆäÀ̽ºº¸´Ù 2.73¹è ´õ ¸¹Àº CPU ÀÚ¿øÀ» »ç¿ëÇÏÁö¸¸, batching I/O Å©±â¿¡ µû¶ó ÃÖ´ë 2.42¹è ³ôÀº IOPS¸¦ º¸¿©ÁÖ¾ú´Ù. ´õºÒ¾î Å°-¹ë·ù ½ºÅä¾î¿¡¼­ÀÇ ½ÇÇè °á°ú, Àбâ À§ÁÖÀÇ ¿öÅ©·Îµå¿¡¼­ ºñµ¿±âÀû I/O ÀÎÅÍÆäÀ̽º¸¦ »ç¿ëÇÑ °æ¿ì, µ¿±âÀû I/O ÀÎÅÍÆäÀ̽º¸¦ »ç¿ëÇÑ °æ¿ì ´ëºñ ÃÖ´ë 6.60¹è ³ôÀº 󸮷®°ú 61.09% ³·Àº ÀÀ´ä½Ã°£À» º¸ÀÌ´Â °ÍÀ» È®ÀÎÇÏ¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
Recently, to utilize high-performance storage devices, various I/O interfaces have been studied. In particular, asynchronous I/O interfaces are now getting attention. However, those asynchronous I/O interfaces have the disadvantage of relatively higher latency and more CPU consumption than synchronous I/O interfaces. In this paper, we first analyze the characteristics of each I/O interface and then apply it to the key-value store to measure the performance at the application layer. The experiments show that the asynchronous I/O interface consumes 2.73¡¿ more CPU resources than the synchronous I/O interface, but delivers up to 2.42¡¿ higher IOPS depending upon the batching I/O size. In addition, experiments on key-value stores show up to 6.60¡¿ higher throughput and 61.09% lower latency when asynchronous I/O interfaces are used in the read-intensive workload.
Å°¿öµå(Keyword) Å°-¹ë·ù ½ºÅä¾î   µ¿±âÀû I/O ó¸® ¹æ½Ä   ºñµ¿±âÀû I/O ó¸® ¹æ½Ä   io_uring ÀÎÅÍÆäÀ̽º   CPU »ç¿ë·®   key-value store   synchronous I/O interface   asynchronous I/O interface   io_uring interface   CPU usage  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå