Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)
ÇѱÛÁ¦¸ñ(Korean Title) |
CPU »ç¿ë·®À» °í·ÁÇÑ °í¼º´É ÀúÀåÀåÄ¡ ±â¹Ý °¡»óÈ ½Ã½ºÅÛÀÇ I/O ¿Ï·á ó¸® ±â¹ý |
¿µ¹®Á¦¸ñ(English Title) |
I/O Completion Technique of Virtualized System Considering CPU Usage with High-Performance Storage Devices |
ÀúÀÚ(Author) |
ÀÌÇýÁö
ÀÌÅÂÇü
À̹ÎÈ£
¼Û¿ëÁÖ
¾ö¿µÀÍ
Hyeji Lee
Taehyung Lee
Minho Lee
Yongju Song
Young Ik Eom
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 46 NO. 07 PP. 0612 ~ 0619 (2019. 07) |
Çѱ۳»¿ë (Korean Abstract) |
ÃÖ±Ù Samsung Z-SSD ¹× Intel Optane SSD¿Í °°Àº °í¼º´É ÀúÀåÀåÄ¡ÀÇ µîÀåÀ¸·Î ÀÎÇØ ½Ã½ºÅÛÀÇ I/O ¼º´É ¿À¹öÇìµå°¡ ÀúÀåÀåÄ¡¿¡¼ ¼ÒÇÁÆ®¿þ¾î I/O °èÃþÀ¸·Î À̵¿ÇÏ¿´´Ù. ÀÌ·Î ÀÎÇØ ÇÏÀÌÆÛ¹ÙÀÌÀú ¹× ¿î¿µÃ¼Á¦´Â °í¼º´É ÀúÀåÀåÄ¡ÀÇ ¼º´ÉÀ» ÃÖ´ëÇÑ È°¿ëÇϱâ À§ÇØ I/O ¿Ï·á ó¸® ±â¹ý Áß ÇϳªÀÎ Æú¸µ ¹æ½ÄÀÇ È¿¿ë¼ºÀ» ÁÖ¸ñÇÏ°í ÀÖÀ¸¸ç, ÇÏÀ̺긮µå Æú¸µ(Hybrid Polling) ¹× ÀûÀÀÇü Æú¸µ(Adaptive Polling)°ú °°Àº »õ·Î¿î ±â¹ýµéÀ» Àû¿ëÇÏ¿© »ç¿ëÇÏ°í ÀÖ´Ù. º» ³í¹®Àº QEMU-KVM ÇÏÀÌÆÛ¹ÙÀÌÀú¿¡¼ Á¦°øÇÏ´Â ÀûÀÀÇü Æú¸µ ¹æ½ÄÀÇ ¹®Á¦Á¡À» ¼³¸íÇÑ µÚ, °í¼º´É ÀúÀåÀåÄ¡ÀÇ ÀÀ´ä ½Ã°£À» ÃÖ´ëÇÑÀ¸·Î È°¿ëÇÏ¸é¼ CPU »ç¿ë·®À» °¨¼Ò½ÃÅ°´Â »õ·Î¿î I/O ¿Ï·á ó¸® ±â¹ýÀ» Á¦¾ÈÇÑ´Ù. ½ÇÇè °á°ú, Á¦¾ÈµÈ ±â¹ýÀº 64KB ÀÌÇÏ Å©±âÀÇ I/O¿äû¿¡ ´ëÇØ ±âÁ¸ ±â¹ý ´ëºñ ÃÖ´ë 5.3% Áö¿¬µÈ ÀÀ´ä ½Ã°£À» º¸¿©ÁÖÁö¸¸, CPU »ç¿ë·®Àº ÃÖ´ë 39.7% °¨¼ÒÇÑ °ÍÀ» È®ÀÎÇÒ ¼ö ÀÖ´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
Recently, the advent of high-performance storage devices such as Samsung Z-SSD and Intel Optane SSD has shifted the I/O systems¡¯ performance overhead from the storage devices to the software I/O layer. To optimize the I/O performance of high-performance storage devices, the hypervisor and operating system have focused on the effectiveness of polling technique, which is one of the I/O completion techniques applied in virtualized systems, and new techniques such as hybrid and adaptive polling are being adopted. This paper reveals the problem of the existing adaptive polling techniques provided by QEMU-KVM hypervisor and proposes a new I/O completion technique, which saves on CPU usage while fully utilizing high-performance storage devices. Our evaluation indicates that the proposed technique reduces CPU usage by up to 39.7% while delaying I/O latency to less than 5.3% only, in comparison to conventional systems.
|
Å°¿öµå(Keyword) |
°¡»óÈ
Æú¸µ
ÀÎÅÍ·´Æ®
°í¼º´É ÀúÀåÀåÄ¡
I/O ¿Ï·á ó¸® ¹æ½Ä
virtualization
polling
interrupt
high-performance storage device
I/O completion method
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|