• 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 :

ÇѱÛÁ¦¸ñ(Korean Title) GPGPU¸¦ ÀÌ¿ëÇÑ À͸íÈ­ ¾Ë°í¸®ÁòÀÇ º´·Äó¸® ±â¹ý
¿µ¹®Á¦¸ñ(English Title) A Parallel Anonymization Algorithm Using General-purpose Graphics Processing Units
ÀúÀÚ(Author) Á¤±âÁ¤   ÀÌÇõ±â   ±è¼öÇü   Á¤¿¬µ·   Kijung Jung   Hyukki Lee   Soohyung Kim   Yon Dohn Chung  
¿ø¹®¼ö·Ïó(Citation) VOL 32 NO. 01 PP. 0088 ~ 0098 (2016. 04)
Çѱ۳»¿ë
(Korean Abstract)
ÃÖ±Ù ´Ù¾çÇÑ ±â°ü°ú ±â¾÷µéÀÌ ¿©·¯ ºÐ¾ß¿¡¼­ ¸¶ÀÌÅ©·Îµ¥ÀÌÅ͸¦ »ý¼ºÇÏ°í À̸¦ ¹èÆ÷ÇÏ°í ÀÖ´Ù. ÀÌ °úÁ¤¿¡¼­ °³ÀÎÀ» ½Äº°ÇÒ ¼ö ÀÖ´Â Á¤º¸¿Í ¹Î°¨ÇÑ Á¤º¸°¡ ÇÔ²² ¹èÆ÷µÇ¾î °³ÀÎÀÇ ÇÁ¶óÀ̹ö½Ã°¡ À¯ÃâµÉ ¼ö ÀÖ´Ù. ¾ÈÀüÇÑ ¸¶ÀÌÅ©·Îµ¥ÀÌÅÍ ¹èÆ÷¸¦ À§ÇÏ¿© k-anonymity¿Í À̸¦ º¸¿ÏÇÑ l-diversity µîÀÇ ÇÁ¶óÀ̹ö½Ã ¸ðµ¨µéÀÌ ³Î¸® È°¿ëµÇ°í ÀÖ´Ù. ÀÌ ¶§, ÇÁ¶óÀ̹ö½Ã ¸ðµ¨À» ¸¸Á·Çϸ鼭 µ¿½Ã¿¡ µ¥ÀÌÅͼÂ(dataset)ÀÇ Á¤º¸ ¼Õ½ÇÀ» ÃÖ¼ÒÈ­ ÇÏ´Â °ÍÀº NP-Hard ¹®Á¦À̱⠶§¹®¿¡, ÀûÀº ½Ã°£ ¾È¿¡ ÃÖÀû¿¡ °¡±î¿î À͸íÈ­ Çظ¦ ãÀ¸·Á´Â ´Ù¾çÇÑ ±Ù»ç ¾Ë°í¸®ÁòµéÀÌ Á¦¾ÈµÇ¾ú´Ù. ±×·¯³ª ºòµ¥ÀÌÅÍ ½Ã´ëÀÇ µµ·¡·Î ÀÎÇØ ¹èÆ÷µÇ´Â µ¥ÀÌÅÍÀÇ Å©±â¿Í ¼Ó¼ºÀÇ ¼ö°¡ Áõ°¡Çϸ鼭 ´õ ºü¸¥ À͸íÈ­ÀÇ Çʿ伺ÀÌ °­Á¶µÇ°í ÀÖ´Ù. µû¶ó¼­ º» ³í¹®¿¡¼­ ±âÁ¸ À͸íÈ­ ¾Ë°í¸®ÁòÀÌ CPU 󸮸¸À» °¡Á¤ÇÑ ¼øÂ÷Àû ¿¬»êÀÓÀ» ÁöÀûÇÏ°í, GPGPU¸¦ ÀÌ¿ëÇÑ À͸íÈ­ ¾Ë°í¸®ÁòÀÇ º´·Ä ó¸® ±â¹ýÀ» Á¦¾ÈÇÑ´Ù. ¶ÇÇÑ, ½ÇÇèÀ» ÅëÇÏ¿© k-anonymity¿Í l-diversity¸¦ ¸¸Á·½ÃÅ°±â À§ÇÑ À͸íÈ­ ¾Ë°í¸®ÁòÀÇ ¼öÇà ½Ã°£À» ºñ±³ÇÏ¿© Á¦¾È ¾Ë°í¸®ÁòÀÇ ¼º´É Çâ»óÀ» °ËÁõÇÑ´Ù.
¿µ¹®³»¿ë
(English Abstract)
In recent years, a number of organizations and companies have created and published microdata in various fields. Data privacy of individuals can be intruded if they are published with sensitive attributes and identifying attributes. As a result, privacy models, such as k-anonymity and l-diversity, are widely used for privacy-preserving microdata publishing. It is an NP-hard problem that satisfying privacy models with minimum information loss. Therefore, varied anonymization algorithms which find sub-optimal answers in less time are suggested. However, due to the advent of the big data era, the number of attributes and the size of datasets increase, hence a necessity for faster anonymization algorithms has been emerged. Thus, we point out the existing algorithms are serial computations which only consider CPU. Also we suggest a parallel anonymization algorithm using general-purpose graphics processing units. Moreover, we demonstrate the performance of the proposed algorithm by comparing with CPU-based algorithm.
Å°¿öµå(Keyword) µ¥ÀÌÅÍ ÇÁ¶óÀ̹ö½Ã   À͸íÈ­   GPGPU   Data privacy   Anonymization   GPGPU  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå